Carbon sink dynamic evaluation and optimization method and system for urban ecology

Abstract

The invention relates to the technical field of urban ecology monitoring and optimization control, and discloses an urban ecology-oriented carbon sink dynamic evaluation and optimization method and system, and the method comprises the steps: obtaining multi-source greenbelt feature data; constructing a light energy utilization rate model to output a photosynthetic carbon fixed quantity sequence; a long-short-term memory network model based on introduction of a weighted trend maintenance loss function and a DropBlock mechanism; setting a situation influence factor to construct a multi-situation evolution result; and generating an optimal greenbelt management regulation and control parameter by adopting a particle swarm algorithm. Compared with the technical problem that high-frequency dynamic carbon sink evaluation and optimization control are difficult to realize for urban different-quality greenbelts especially under the condition of frequent manual management behaviors in the prior art, the method has the advantages that the generalization ability of the model is improved by introducing a trend maintenance loss mechanism and time slice DropBlock regularization, and the multi-situation disturbance simulation is combined, so that the dynamic carbon sink evaluation and optimization control of the urban different-quality greenbelts can be realized, and the dynamic carbon sink evaluation and optimization control of the urban different-quality greenbelts can be realized. And the real-time performance, the accuracy and the regulation and control optimization efficiency of urban carbon sink monitoring are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of urban ecological monitoring and optimization control, and in particular to a carbon sink dynamic assessment and optimization method and system for urban ecology. Background Art

[0002] Traditional carbon sink assessment methods primarily rely on remote sensing imagery and meteorological data to construct vegetation index models (such as NDVI and EVI) or employ light energy utilization models (such as CASA) for estimation. These methods are typically based on fixed timescales, regional average parameters, and static green space assumptions. These methods are ill-suited to the complex structures, diverse management practices, and frequent environmental disturbances typical of urban green spaces. For example, street greenbelts, vertical greening, and park green spaces exhibit significant differences in morphological structure, light intake, temperature and humidity fluctuations, and frequency of human intervention. However, traditional models often lack categorical modeling mechanisms and are unable to dynamically respond to these heterogeneous characteristics. Furthermore, existing models generally rely on annual or monthly average data, failing to achieve real-time updates at daily or even higher frequencies, and thus failing to predict carbon sink trends. Furthermore, current urban green space management practices (such as irrigation, pruning, and replanting) significantly impact carbon sequestration capacity, but existing technologies rarely incorporate these management practices as modeling variables into prediction frameworks, resulting in significant discrepancies between predicted and actual results. Furthermore, existing technologies generally lack effective feedback optimization mechanisms, making it difficult to dynamically generate management strategies based on carbon sink prediction results. Therefore, there is an urgent need for a carbon sink assessment and optimization method that integrates urban heterogeneous green spaces, multi-source data fusion, high-frequency dynamic prediction, and intelligent regulation, which can achieve accurate assessment of carbon sink changes, trend prediction, and strategy optimization even when green space types are complex and human intervention is frequent. Summary of the Invention

[0003] In response to the above-mentioned technical deficiencies, the purpose of the present invention is to propose a dynamic carbon sink assessment and optimization method for urban ecology, aiming to solve the technical problem in the existing technology that it is difficult to achieve high-frequency dynamic carbon sink assessment and optimization control for urban green spaces of different qualities, especially under conditions of frequent artificial management.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a carbon sink dynamic assessment and optimization method for urban ecology.

[0005] The carbon sink dynamic assessment and optimization method for urban ecology includes:

[0006] Step S10: Obtain the time of day through the preset multi-source sensor network and remote sensing data source Green space feature dataset , green space feature dataset Including NDVI value, temperature , soil moisture , light intensity , management behavior variables and time index ;

[0007] Step S20: Construct a light energy utilization model and convert the green space feature dataset As the input of the light energy utilization model, the output is the photosynthetic carbon fixation sequence ;

[0008] Step S30: Sequence photosynthetic carbon fixation As input to a pre-trained long-short-term memory network, the output is a predicted sequence of photosynthetic carbon fixation for the next day. A weighted trend-preserving loss function is introduced during the pre-training of the long-short-term memory network, and a DropBlock mechanism is used to prevent overfitting.

[0009] Step S40: obtaining the current green space type, setting a scenario impact factor according to the current green space type, mapping the scenario impact factor to the photosynthetic carbon fixation prediction sequence, and performing cluster analysis to obtain a multi-scenario evolution result;

[0010] Step S50: A multi-objective function for carbon sink optimization is set based on the multi-scenario evolution results, and a particle swarm algorithm is used to solve the multi-objective function for carbon sink optimization. Each particle represents a set of candidate control strategies u(t). By updating the particle position and velocity and iteratively approaching the optimal solution, the optimal management and control parameters for each green space type in the next day are obtained.

[0011] Preferably, in step S40, the current green space type includes 、 and ; Scenario influencing factors include temperature rise influencing factors, precipitation change influencing factors, management frequency influencing factors and green space expansion influencing factors.

[0012] Preferably, in step S20, the light energy utilization efficiency model includes a three-layer shallow structure, specifically including:

[0013] The first layer is the radiation absorption estimation layer, which is used to extract the actual absorbed photosynthetic active radiation characteristics based on NDVI and light intensity;

[0014] The second layer, the environmental adaptation control layer, is used to extract light energy utilization efficiency characteristics based on temperature, soil moisture, and management behavior variables;

[0015] The third layer, the carbon sink output calculation layer, is used to perform product-integral calculations based on the characteristics of photosynthetic active radiation and the characteristics of extracted light energy utilization efficiency, and output a photosynthetic carbon fixation sequence in the form of a time series in combination with the time index.

[0016] Preferably, in step S30, the step of introducing a weighted trend preserving loss function in the pre-training process of the long short-term memory network specifically includes:

[0017] Construct training sample pairs based on photosynthetic carbon fixation time series and green space characteristic dataset ,in, Time of day forward Green space feature dataset of the day, For the The output of the time series of photosynthetic carbon fixation will be As a series of predicted carbon sink values ;

[0018] Construct the corresponding real carbon sink value sequence based on the preset historical data , according to the real carbon sink value series and the predicted value series Construct a weighted trend preservation loss function. The formula used by the weighted trend preservation loss function is:

[0019] ;

[0020] in, Maintain loss function for weight trend; , ; is the trend constraint weight coefficient, which is used to control the sensitivity to trend changes; Time of day The real carbon sink value series, No. The real carbon sink value series of the day, Time of day The predicted carbon sink value series, No. The predicted carbon sink value series of the day; ( ) is the mean square error function between the true trend change and the predicted trend change.

[0021] Preferably, in step S30, the step of using the DropBlock mechanism to prevent overfitting during the pre-training process of the long short-term memory network specifically includes:

[0022] Two parameters of the DropBlock mechanism are pre-set, including: the masking probability p, which is used to indicate the probability of applying the DropBlock mechanism in each training batch; the masking block length , which is used to represent the number of consecutive masked time steps in the time series dimension;

[0023] For each green space feature dataset in the training sample pair Sample a starting position in its time dimension , the construction length is the shield block length Mask window , and construct the mask matrix , where the i-th item in the mask matrix is ; The mask matrix Acting on green space feature dataset Obtain a masked green space feature dataset to prevent overfitting , ,in Represents element-wise multiplication.

[0024] Preferably, in step S40, the multi-scenario evolution results include a carbon sink response elasticity index, a sensitive factor contribution ranking table and a green space type scenario adaptability map; wherein, the carbon sink response elasticity index includes the absolute response amplitude, the relative carbon sink change rate and the change trend offset; the sensitive factor contribution ranking table is obtained by passing a single scenario influencing factor and keeping other factors unchanged, and is used to analyze the marginal changes in carbon sink output; the green space type scenario adaptability map is used to support green space priority classification.

[0025] Preferably, in step S50, the carbon sink optimization multi-objective function includes a carbon sink compliance objective function and a control behavior cost objective function; the selected control strategy u(t) is set through a remote cloud platform, and the influence weight of the selected control strategy u(t) on the action space in the particle swarm algorithm is pre-set by expert experience; the optimal management and control parameters include the optimal irrigation frequency, the optimal pruning frequency and the optimal green space deployment personnel configuration.

[0026] The present invention also provides a carbon sink dynamic assessment and optimization system for urban ecology, including:

[0027] Green space feature acquisition module, used to obtain the time of day through the preset multi-source sensor network and remote sensing data source Green space feature dataset , green space feature dataset Including NDVI value, temperature , soil moisture , light intensity , management behavior variables and time index ;

[0028] Light energy utilization modeling module is used to build a light energy utilization model and integrate the green space feature dataset As the input of the light energy utilization model, the output is the photosynthetic carbon fixation sequence ;

[0029] Carbon sink prediction module, used to convert photosynthetic carbon fixation sequence As input to a pre-trained long-short-term memory network, the output is a predicted sequence of photosynthetic carbon fixation for the next day. A weighted trend-preserving loss function is introduced during the pre-training of the long-short-term memory network, and a DropBlock mechanism is used to prevent overfitting.

[0030] The multi-scenario evolution simulation module is used to obtain the current green space type, set the scenario impact factor according to the current green space type, map the scenario impact factor to the photosynthetic carbon fixation prediction sequence, and obtain the multi-scenario evolution results;

[0031] The optimization and control strategy generation module is used to set the carbon sink optimization multi-objective function based on the multi-scenario evolution results, and adopts the particle swarm algorithm to solve the carbon sink optimization multi-objective function. Each particle represents a set of candidate control strategies u(t). By updating the particle position and velocity, the optimal solution is iterated and the optimal management and control parameters of each green space type in the next day are obtained.

[0032] The present invention also provides a carbon sink dynamic assessment and optimization device for urban ecology, comprising: a memory, a processor, and a carbon sink dynamic assessment and optimization program for urban ecology stored in the memory and runnable on the processor. When the carbon sink dynamic assessment and optimization program for urban ecology is executed by the processor, a carbon sink dynamic assessment and optimization method for urban ecology is implemented.

[0033] The present invention also provides a computer program product, including a carbon sink dynamic assessment and optimization program for urban ecology. When the carbon sink dynamic assessment and optimization program for urban ecology is executed by a processor, the carbon sink dynamic assessment and optimization method for urban ecology is implemented.

[0034] The beneficial effect of the present invention is that compared with the existing technology, especially under conditions of frequent manual management, it is difficult to achieve high-frequency dynamic carbon sink assessment and optimization control for urban green spaces of different qualities. This application improves the generalization ability of the model by introducing a trend-keeping loss mechanism and time segment DropBlock regularization, and combines it with multi-scenario disturbance simulation to improve the real-time, accuracy and regulation optimization efficiency of urban carbon sink monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1This is a flow chart of a method for dynamic carbon sequestration assessment and optimization for urban ecology according to the present invention.

[0037] Figure 2 This is a structural diagram of a carbon sink dynamic assessment and optimization device for urban ecology according to the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] Example 1: Figure 1 , which is a flow chart of the method for dynamic assessment and optimization of carbon sinks for urban ecology according to the present invention, and provides Example 1 of the method for dynamic assessment and optimization of carbon sinks for urban ecology according to the present invention.

[0040] In the first embodiment, the urban ecological carbon sink dynamic assessment and optimization method includes:

[0041] Step S10: Obtain the time of day through the preset multi-source sensor network and remote sensing data source Green space feature dataset , green space feature dataset Including NDVI value, temperature , soil moisture , light intensity , management behavior variables and time index ;

[0042] It should be noted that the multi-source sensor network may include micro-meteorological sensors, soil moisture monitoring nodes, light intensity acquisition modules, and personnel input terminals for collecting management activity information installed around the green space. The remote sensing data source can use multispectral satellite images with a resolution better than 10 meters (such as Sentinel-2) to obtain daily NDVI vegetation index values ​​and spatially align the green space locations through the urban geographic information system (GIS). The collection of green space characteristic data does not rely on a single device or a single moment, but rather integrates sources at different time points, different scales, and different data types to form a structured data set in a unified format to provide a data basis for subsequent carbon sink modeling.

[0043] It should be understood that management behavior variables include not only explicit operational activities such as irrigation, pruning, and fertilization, but can also be quantified through log records, automated work orders, or maintenance frequency estimates to represent the intensity of human intervention. The introduction of such variables significantly enhances the model's ability to model non-ecological factors.

[0044] Step S20: Construct a light energy utilization model and convert the green space feature dataset As the input of the light energy utilization model, the output is the photosynthetic carbon fixation sequence ;

[0045] It should be noted that in step S20, the light energy utilization model includes three shallow structures, specifically including: the first layer of radiation absorption estimation layer, which is used to extract the actual absorbed photosynthetic active radiation characteristics based on NDVI and light intensity; the second layer of environmental adaptation control layer, which is used to extract light energy utilization efficiency characteristics based on air temperature, soil moisture, and management behavior variables; the third layer of carbon sink output calculation layer, which is used to perform product-integral calculations based on photosynthetic active radiation characteristics and extracted light energy utilization efficiency characteristics, and combine the time index to output a photosynthetic carbon fixation sequence in the form of a time series.

[0046] It's understandable that light intensity and NDVI primarily determine the basic photosynthetic capacity of vegetation, while temperature, humidity, and human management practices influence the ultimate carbon fixation by regulating photosynthetic efficiency. Therefore, the model extracts light energy utilization efficiency characteristics in the second layer based on temperature, soil moisture, and management behavior variables. This light energy utilization efficiency model is not only applicable to conventional urban greening units such as street greenbelts and park green spaces, but can also be extended to special green spaces such as vertical greening and rooftop greening.

[0047] Step S30: Sequence photosynthetic carbon fixation As input to a pre-trained long-short-term memory network, the output is a predicted sequence of photosynthetic carbon fixation for the next day. A weighted trend-preserving loss function is introduced during the pre-training of the long-short-term memory network, and a DropBlock mechanism is used to prevent overfitting.

[0048] It should be noted that in step S30, the step of introducing the weighted trend preservation loss function in the pre-training process of the long short-term memory network specifically includes: constructing a training sample pair based on the photosynthetic carbon fixation time series and the green space feature dataset. ,in, Time of day forward Green space feature dataset of the day, For the The output of the time series of photosynthetic carbon fixation will be As a series of predicted carbon sink values ; Construct the corresponding real carbon sink value sequence based on the preset historical data , according to the real carbon sink value series and the predicted value series Construct a weighted trend preservation loss function. The formula used by the weighted trend preservation loss function is: ,in, Maintain loss function for weight trend; , ; is the trend constraint weight coefficient, which is used to control the sensitivity to trend changes; Time of day The real carbon sink value series, No. The real carbon sink value series of the day, Time of day The predicted carbon sink value series, No. The predicted carbon sink value series of the day; ( ) is the mean square error function between the actual trend change and the predicted trend change. In step S30, the DropBlock mechanism is used to prevent overfitting during the pre-training of the long short-term memory network, specifically including: pre-setting two parameters of the DropBlock mechanism, including: the masking probability p, which is used to represent the probability of applying the DropBlock mechanism in each training batch; the masking block length , used to represent the number of consecutive masked time steps in the time series dimension; for each green space feature dataset in the training sample pair Sample a starting position in its time dimension , the construction length is the shield block length Mask window , and construct the mask matrix , where the i-th item in the mask matrix is ; The mask matrix Acting on green space feature dataset Obtain a masked green space feature dataset to prevent overfitting , ,in Represents element-wise multiplication.

[0049] As you can see, the weighted trend-preserving loss function not only penalizes prediction errors but also explicitly considers the accuracy of carbon sink growth or decline trends; this is particularly important for short-term predictions of urban green spaces under sudden climate changes and unexpected maintenance behaviors. Furthermore, the DropBlock mechanism randomly blocks consecutive time segments to prevent the model from overfitting to inputs from a single time period, thereby improving generalization performance.

[0050] It should be understood that compared with the traditional LSTM training method that only uses the MSE loss function or the ordinary Dropout mechanism, the training mechanism constructed in this embodiment can better enhance the model's learning ability of complex carbon sink dynamics, especially when the performance of different types of green spaces in cities varies greatly and the environment fluctuates strongly, it can still maintain a high trend consistency and prediction accuracy.

[0051] Step S40: obtaining the current green space type, setting a scenario impact factor according to the current green space type, mapping the scenario impact factor to the photosynthetic carbon fixation prediction sequence, and performing cluster analysis to obtain a multi-scenario evolution result;

[0052] It should be noted that in step S40, the multi-scenario evolution results include the carbon sink response elasticity index, the sensitive factor contribution ranking table and the green space type scenario adaptability map; among them, the carbon sink response elasticity index includes the absolute response amplitude, the relative carbon sink change rate and the change trend offset; the sensitive factor contribution ranking table is obtained by passing a single scenario influencing factor and keeping other factors unchanged, which is used to analyze the marginal changes in carbon sink output; the green space type scenario adaptability map is used to support the green space priority classification.

[0053] It is understandable that through scenario simulation and response indicator extraction, not only can the trend of carbon sink changes under the current management plan be identified, but also the potential impact of future climate change or policy intervention on the carbon sink system can be assessed in advance.

[0054] It should be understood that the present invention is not limited to the prediction of static carbon sink values, but through multi-scenario analysis, obtains the dynamic behavior mapping of the carbon sink system under different external disturbances, thereby providing input-based management strategies for subsequent optimization rather than relying solely on empirical rules, significantly improving the scientific nature and adaptability of the control strategy.

[0055] Step S50: A multi-objective function for carbon sink optimization is set based on the multi-scenario evolution results, and a particle swarm algorithm is used to solve the multi-objective function for carbon sink optimization. Each particle represents a set of candidate control strategies u(t). By updating the particle position and velocity and iteratively approaching the optimal solution, the optimal management and control parameters for each green space type in the next day are obtained.

[0056] It should be noted that in step S50, the carbon sink optimization multi-objective function includes the carbon sink compliance objective function and the control behavior cost objective function; the selected control strategy u(t) is set through the remote cloud platform, and the influence weight of the selected control strategy u(t) on the action space in the particle swarm algorithm is pre-set by expert experience; the optimal management and control parameters include the optimal irrigation frequency, the optimal pruning frequency and the optimal green space deployment personnel configuration.

[0057] It's understandable that incorporating expert knowledge into the control variable weighting can avoid carbon sink deviations and resource waste caused by blind searches. Furthermore, the multi-scenario weighting approach makes the resulting strategy more robust, maintaining stable carbon sink performance despite climate disturbances and data uncertainty.

[0058] It should be understood that this optimization strategy not only outputs a single optimal solution, but also forms a set of strategy alternatives, which can be deployed in different green space management systems according to budget, cycle or scheduling priority. It is suitable for application scenarios such as hierarchical deployment, regional coordination, and rolling updates, and has strong engineering expansion value.

[0059] Embodiment 2: In addition, the present invention provides a carbon sink dynamic assessment and optimization system for urban ecology, which adopts a carbon sink dynamic assessment and optimization method for urban ecology in the above embodiment, and can solve the technical problem of carbon sink dynamic assessment and optimization for urban ecology. Compared with the existing technology, the beneficial effects of the carbon sink dynamic assessment and optimization system for urban ecology provided by the present invention are the same as the beneficial effects of the carbon sink dynamic assessment and optimization method for urban ecology provided by the above embodiment, and the other technical features of the carbon sink dynamic assessment and optimization system for urban ecology are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0060] Example 3: The present invention provides a carbon sink dynamic assessment and optimization device for urban ecology, please refer to Figure 2A device for dynamic carbon sink assessment and optimization for urban ecology includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method for dynamic carbon sink assessment and optimization for urban ecology described in the first embodiment. The device for dynamic carbon sink assessment and optimization for urban ecology in this embodiment of the present invention may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. The device for dynamic carbon sink assessment and optimization for urban ecology is merely an example and should not limit the functionality or scope of use of this embodiment of the present invention. A device for dynamic carbon sequestration assessment and optimization for urban ecology may include a processor 1001 (e.g., a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of the device. Processor 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007, such as a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008, such as a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003, such as a magnetic tape, hard disk, etc.; and communication devices 1009. Communication devices 1009 can allow the urban ecology-oriented carbon sink dynamic assessment and optimization device to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a carbon sink dynamic assessment and optimization device for urban ecology with various systems, it should be understood that implementation or presence of all the illustrated systems is not required. More or fewer systems may alternatively be implemented or present.

[0061] Example 4: The present invention also provides a computer program product, comprising a computer program. When executed by a processor, the computer program implements the steps of the above-described method for dynamic carbon sink assessment and optimization for urban ecology. The computer program product provided by the present invention can solve the technical problem of dynamic carbon sink assessment and optimization for urban ecology. Compared with the prior art, the beneficial effects of the computer program product provided by the present invention are the same as those of the method for dynamic carbon sink assessment and optimization for urban ecology provided in the above-described embodiment, and are not further described here.

[0062] In particular, according to the embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processor 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present invention are performed.

[0063] It should be understood that the various parts disclosed in the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any appropriate manner in any one or more embodiments or examples.

[0064] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for dynamic carbon sink assessment and optimization for urban ecology, characterized by: Methods include: Step S10: Obtain the time of day through the preset multi-source sensor network and remote sensing data source Green space feature dataset , green space feature dataset Including NDVI value, temperature , soil moisture , light intensity , management behavior variables and time index ; Step S20: Construct a light energy utilization model and convert the green space feature dataset As the input of the light energy utilization model, the output is the photosynthetic carbon fixation sequence ; Step S30: Sequence photosynthetic carbon fixation As input to a pre-trained long-short-term memory network, the output is a predicted sequence of photosynthetic carbon fixation for the next day. A weighted trend-preserving loss function is introduced during the pre-training of the long-short-term memory network, and a DropBlock mechanism is used to prevent overfitting. Step S40: obtaining the current green space type, setting a scenario impact factor according to the current green space type, mapping the scenario impact factor to the photosynthetic carbon fixation prediction sequence, and performing cluster analysis to obtain a multi-scenario evolution result; Step S50: A multi-objective function for carbon sink optimization is set based on the multi-scenario evolution results, and a particle swarm algorithm is used to solve the multi-objective function for carbon sink optimization. Each particle represents a set of candidate control strategies u(t). By updating the particle position and velocity and iteratively approaching the optimal solution, the optimal management and control parameters for each green space type in the next day are obtained.

2. The method for dynamic carbon sink assessment and optimization for urban ecology according to claim 1, characterized in that: In step S40, the current green space type includes 、 and ; Scenario influencing factors include temperature increase influencing factors, precipitation change influencing factors, management frequency influencing factors and green space expansion influencing factors.

3. The method for dynamic carbon sink assessment and optimization for urban ecology according to claim 1, characterized in that: In step S20, the light energy utilization model includes a three-layer shallow structure, specifically including: The first layer is the radiation absorption estimation layer, which is used to extract the actual absorbed photosynthetic active radiation characteristics based on NDVI and light intensity; The second layer, the environmental adaptation control layer, is used to extract light energy utilization efficiency characteristics based on temperature, soil moisture, and management behavior variables; The third layer, the carbon sink output calculation layer, is used to perform product-integral calculations based on the characteristics of photosynthetic active radiation and the characteristics of extracted light energy utilization efficiency, and output a photosynthetic carbon fixation sequence in the form of a time series in combination with the time index.

4. The method for dynamic carbon sink assessment and optimization for urban ecology according to claim 1, characterized in that: In step S30, the step of introducing a weighted trend-preserving loss function during the pre-training process of the long short-term memory network specifically includes: Construct training sample pairs based on photosynthetic carbon fixation time series and green space characteristic dataset ,in, Time of day forward Green space feature dataset of the day, For the The output of the time series of photosynthetic carbon fixation will be As a series of predicted carbon sink values ; Construct the corresponding real carbon sink value sequence based on the preset historical data , according to the real carbon sink value series and the predicted value series Construct a weighted trend preservation loss function. The formula used by the weighted trend preservation loss function is: ； in, Maintain loss function for weight trend; , ; is the trend constraint weight coefficient, which is used to control the sensitivity to trend changes; Time of day The real carbon sink value series, No. The real carbon sink value series of the day, Time of day The predicted carbon sink value series, No. The predicted carbon sink value series of the day; ( ) is the mean square error function between the true trend change and the predicted trend change.

5. The method for dynamic carbon sink assessment and optimization for urban ecology according to claim 4, characterized in that: In step S30, the steps of using the DropBlock mechanism to prevent overfitting during the pre-training of the long short-term memory network include: Two parameters of the DropBlock mechanism are pre-set, including: the masking probability p, which is used to indicate the probability of applying the DropBlock mechanism in each training batch; the masking block length , which is used to represent the number of consecutive masked time steps in the time series dimension; For each green space feature dataset in the training sample pair Sample a starting position in its time dimension , the construction length is the shield block length Mask window , and construct the mask matrix , where the i-th item in the mask matrix is ; The mask matrix Acting on green space feature dataset Obtain a masked green space feature dataset to prevent overfitting , ,in Represents element-wise multiplication.

6. The method for dynamic carbon sink assessment and optimization for urban ecology according to claim 1, characterized in that: In step S40, the multi-scenario evolution results include a carbon sink response elasticity index, a sensitive factor contribution ranking table, and a green space type scenario adaptability map; wherein, the carbon sink response elasticity index includes the absolute response amplitude, the relative carbon sink change rate, and the change trend offset; the sensitive factor contribution ranking table is obtained by influencing a single scenario factor and keeping other factors unchanged, and is used to analyze the marginal changes in carbon sink output; the green space type scenario adaptability map is used to support green space priority classification.

7. The method for dynamic carbon sink assessment and optimization for urban ecology according to claim 1, characterized in that: In step S50, the carbon sink optimization multi-objective function includes a carbon sink compliance objective function and a control behavior cost objective function; the selected control strategy u(t) is set through a remote cloud platform, and the influence weight of the selected control strategy u(t) on the action space in the particle swarm algorithm is pre-set by expert experience; the optimal management and control parameters include the optimal irrigation frequency, the optimal pruning frequency and the optimal green space deployment personnel configuration.

8. A carbon sink dynamic assessment and optimization system for urban ecology, applied to a carbon sink dynamic assessment and optimization method for urban ecology according to any one of claims 1 to 7, characterized in that: The urban ecological carbon sink dynamic assessment and optimization system includes: Green space feature acquisition module, used to obtain the time of day through the preset multi-source sensor network and remote sensing data source Green space feature dataset , green space feature dataset Including NDVI value, temperature , soil moisture , light intensity , management behavior variables and time index ; Light energy utilization modeling module is used to build a light energy utilization model and integrate the green space feature dataset As the input of the light energy utilization model, the output is the photosynthetic carbon fixation sequence ; Carbon sink prediction module, used to convert photosynthetic carbon fixation sequence As input to a pre-trained long-short-term memory network, the output is a predicted sequence of photosynthetic carbon fixation for the next day. A weighted trend-preserving loss function is introduced during the pre-training of the long-short-term memory network, and a DropBlock mechanism is used to prevent overfitting. The multi-scenario evolution simulation module is used to obtain the current green space type, set the scenario impact factor according to the current green space type, map the scenario impact factor to the photosynthetic carbon fixation prediction sequence, and obtain the multi-scenario evolution results; The optimization and control strategy generation module is used to set the carbon sink optimization multi-objective function based on the multi-scenario evolution results, and adopts the particle swarm algorithm to solve the carbon sink optimization multi-objective function. Each particle represents a set of candidate control strategies u(t). By updating the particle position and velocity, the optimal solution is iterated and the optimal management and control parameters of each green space type in the next day are obtained.

9. A carbon sink dynamic assessment and optimization device for urban ecology, characterized by: The carbon sink dynamic assessment and optimization device for urban ecology includes: a memory, a processor, and a carbon sink dynamic assessment and optimization program for urban ecology stored on the memory and runnable on the processor. When the carbon sink dynamic assessment and optimization program for urban ecology is executed by the processor, a carbon sink dynamic assessment and optimization method for urban ecology according to any one of claims 1 to 7 is implemented.

10. A computer program product, characterized in that The computer program product includes a carbon sink dynamic assessment and optimization program for urban ecology. When the carbon sink dynamic assessment and optimization program for urban ecology is executed by a processor, it implements a carbon sink dynamic assessment and optimization method for urban ecology according to any one of claims 1 to 7.

Images