Plant maintenance method, system, device and storage medium based on traffic pollution

Abstract

The application discloses a plant maintenance method, system and device based on traffic pollution and a storage medium, and the technical scheme points of the application are as follows: traffic congestion data, meteorological data, air pollution data, soil physicochemical data and plant physiological data of target plants in a target area are acquired; a pollution diffusion coefficient is calculated according to the meteorological data and the air pollution data; a first composite feature is constructed according to the pollution diffusion coefficient, the meteorological data and the air pollution data, a second composite feature is constructed according to the pollution diffusion coefficient, the traffic congestion data and the meteorological data, and a third composite feature is constructed according to the soil physicochemical data and the plant physiological data; the first composite feature, the second composite feature and the third composite feature are input into a pre-trained pollution prediction model to obtain a pollution prediction result; and a plant maintenance system corresponding to the target plants is controlled according to the pollution prediction result. The application helps to improve the ability of the target plants to resist urban traffic pollution and relieve the degree of oxidative damage.

Description

Technical Field

[0001] This invention belongs to the field of plant maintenance technology, specifically relating to a plant maintenance method, system, equipment, and storage medium based on traffic pollution. Background Technology

[0002] Current methods for plant care typically involve timed and measured irrigation or fertilization. However, with accelerating urbanization and increasing traffic pollution, plants suffer oxidative damage, impacting their effectiveness in urban ecological restoration.

[0003] Existing plant care methods do not take into account the impact of traffic pollution on the real-time oxidative stress state of plants, resulting in limited effectiveness in improving plant stress resistance. Summary of the Invention

[0004] The purpose of this invention is to provide a plant care method, system, equipment, and storage medium based on traffic pollution, which helps to improve the target plant's tolerance to urban traffic pollution, alleviate its oxidative damage, and enhance the plant's stress resistance.

[0005] The first aspect of this invention provides a plant maintenance method based on traffic pollution, comprising:

[0006] Acquire traffic congestion data, meteorological data, air pollution data, soil physicochemical data, and plant physiological data of target plants in the target area, wherein the target plants are planted in the target area;

[0007] The pollution diffusion coefficient was calculated based on the meteorological and air pollution data.

[0008] A first composite feature is constructed based on the pollution diffusion coefficient, meteorological data, and air pollution data; a second composite feature is constructed based on the pollution diffusion coefficient, traffic congestion data, and meteorological data; and a third composite feature is constructed based on the soil physicochemical data and plant physiological data.

[0009] The first composite feature, the second composite feature, and the third composite feature are input into a pre-trained pollution prediction model to obtain pollution prediction results.

[0010] The plant maintenance system corresponding to the target plant is controlled based on the pollution prediction results.

[0011] In some embodiments, calculating the pollution diffusion coefficient based on the meteorological data and air pollution data includes:

[0012] Pollution gradient data were calculated based on the air pollution data.

[0013] The pollution diffusion coefficient is calculated based on the pollution gradient data and the wind speed in the meteorological data.

[0014] In some implementations, constructing the first composite feature based on the pollution diffusion coefficient, meteorological data, and air pollution data includes:

[0015] Wind speed and temperature sequences are extracted from meteorological data based on a preset sliding time window.

[0016] The pollution diffusion sequence is extracted from the pollution diffusion coefficient based on the sliding time window.

[0017] The pollution concentration sequence is extracted from the air pollution data based on the sliding time window.

[0018] The pollution diffusion sequence, wind speed sequence, temperature sequence, and pollution concentration sequence are combined to obtain the first composite feature.

[0019] In some implementations, constructing the second composite feature based on the pollution diffusion coefficient, traffic congestion data, and meteorological data includes:

[0020] The pollution diffusion sequence is extracted from the pollution diffusion coefficient based on a preset sliding time window;

[0021] Emission intensity data is calculated based on the traffic congestion data, and an emission intensity sequence is extracted from the emission intensity data based on the sliding time window.

[0022] Wind speed sequences are extracted from meteorological data based on the sliding time window.

[0023] The wind speed sequence, pollution diffusion sequence, and emission intensity sequence are combined to obtain the second composite feature.

[0024] In some embodiments, constructing the third composite feature based on the soil physicochemical data and plant physiological data includes:

[0025] The bioavailability of heavy metals was calculated based on the soil physicochemical data, and the plant stress resistance index and plant purification potential of the target plant were calculated based on the plant physiological data.

[0026] The ecological buffering capacity is obtained by calculating the geometric mean of the bioavailability, plant stress resistance index and plant purification potential.

[0027] The stress resistance index sequence is extracted from the plant stress resistance index according to the preset sliding time window, the buffer capacity sequence is extracted from the ecological buffer capacity according to the sliding time window, the bioavailability is replicated and extended to the time step dimension to obtain the bioavailability sequence, and the plant purification potential is replicated and extended to the time step dimension to obtain the purification potential sequence.

[0028] The bioavailability sequence, plant stress resistance index sequence, plant purification potential sequence, and buffering capacity sequence are combined to obtain the third composite feature.

[0029] In some implementations, the pollution prediction model is constructed based on a long short-term memory network, and the pollution prediction model includes:

[0030] The first channel constructed by the two-layer long short-term memory network is used to perform temporal analysis on the first composite feature of the input to obtain the first hidden state vector;

[0031] The second channel constructed by the single-layer long short-term memory network is used to perform feature concatenation and analysis on the input second composite feature and third composite feature to obtain the second hidden state vector.

[0032] The feature fusion layer is used to perform weighted fusion of the first hidden state vector and the second hidden state vector to obtain a weighted fusion vector.

[0033] The output layer is used to decode and temporally expand the weighted fusion vector to obtain the pollution prediction result.

[0034] In some implementations, the first channel includes:

[0035] A first input layer is used to receive the first composite feature;

[0036] The first long short-term memory network layer is used to extract the hidden state vector of the first composite feature at each time step;

[0037] The second long short-term memory network layer is used to aggregate the hidden state vectors of all time steps across time steps to obtain the first hidden state vector of the final time step.

[0038] The first output layer is used to output the first hidden state vector;

[0039] The second channel includes:

[0040] The second input layer is used to receive the second composite feature and the third composite feature;

[0041] The feature splicing layer is used to splice the second composite feature and the third composite feature to obtain the spliced ​​feature;

[0042] The third long short-term memory network layer is used to aggregate the spliced ​​features across time steps to obtain the second hidden state vector of the final time step.

[0043] The output layer is used to output the second hidden state vector.

[0044] A second aspect of the present invention provides a plant maintenance system based on traffic pollution, comprising:

[0045] The data acquisition module is used to acquire traffic congestion data, meteorological data, air pollution data, soil physicochemical data, and plant physiological data of target plants in the target area, wherein the target plants are planted in the target area;

[0046] The calculation module is used to calculate the pollution diffusion coefficient based on the meteorological data and air pollution data;

[0047] The feature construction module is used to construct a first composite feature based on the pollution diffusion coefficient, meteorological data and air pollution data, construct a second composite feature based on the pollution diffusion coefficient, traffic congestion data and meteorological data, and construct a third composite feature based on soil physicochemical data and plant physiological data.

[0048] The pollution prediction module is used to input the first composite feature, the second composite feature and the third composite feature into a pre-trained pollution prediction model to obtain pollution prediction results.

[0049] The maintenance control module is used to control the plant maintenance system corresponding to the target plant based on the pollution prediction results.

[0050] A third aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0051] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described above.

[0052] The technical solution provided by this invention has the following advantages and effects: by combining traffic congestion data, meteorological data, soil physicochemical data and plant physiological data to predict the pollution results of the target area, it is convenient to carry out maintenance treatment on the target plants in the target area in advance according to the pollution prediction results, which helps to improve the target plants' ability to tolerate urban traffic pollution and alleviate their degree of oxidative damage. Attached Figure Description

[0053] Figure 1 This is a flowchart illustrating the plant maintenance method based on traffic pollution provided by the present invention.

[0054] Figure 2 This is a data graph of POD, SOD, and CAT from the experimental results of the first set of experiments provided by this invention;

[0055] Figure 3 This is a data graph of MDA and PRO from the experimental results of the first set of experiments provided by this invention;

[0056] Figure 4 This is a data graph of POD, SOD, and CAT from the experimental results of the second set of experiments provided by this invention;

[0057] Figure 5 This is a data graph of MDA and PRO from the experimental results of the second set of experiments provided by this invention;

[0058] Figure 6 This is a structural block diagram of the plant maintenance system based on traffic pollution provided by the present invention;

[0059] Figure 7 This is an internal structural diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0060] To facilitate understanding of the present invention, specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings.

[0061] Unless otherwise specified or defined, the terms "first," "second," etc., used in this document are for distinguishing names only and do not represent a specific number or order.

[0062] Unless otherwise stated or defined, the term “and / or” as used herein includes any and all combinations of one or more of the related listed items.

[0063] It should be noted that in this article, "fixed to" or "connected to" can mean directly fixed to or connected to a component, or indirectly fixed to or connected to a component.

[0064] like Figure 1 As shown, this embodiment provides a plant maintenance method based on traffic pollution, including the following steps S1~S5:

[0065] Step S1: Obtain traffic congestion data, meteorological data, air pollution data, soil physicochemical data, and plant physiological data of the target plant in the target area, wherein the target plant is planted in the target area.

[0066] In practical applications, traffic congestion data for the target area can be obtained by accessing the Gaode Map API. For example, traffic congestion delay index and vehicle speed can be obtained by accessing the Gaode Map API, and the traffic congestion delay index can be used as traffic congestion data. Collected meteorological data includes wind speed and temperature; collected air pollution data includes PM2.5 concentration, NO2 concentration, SO2 concentration, etc.; collected soil physicochemical data includes extractable metal content, total heavy metal content, and soil buffering capacity, etc.; collected plant physiological data includes leaf antioxidant enzyme activity and oxidative damage marker content, etc.

[0067] Step S2: Calculate the pollution diffusion coefficient based on the meteorological data and air pollution data.

[0068] Specifically, the calculation of the pollution diffusion coefficient based on the meteorological data and air pollution data includes:

[0069] Pollution gradient data were calculated based on the air pollution data.

[0070] The pollution diffusion coefficient is calculated based on the pollution gradient data and the wind speed in the meteorological data.

[0071] In practical applications, the air pollution concentrations at a first monitoring point and a second monitoring point in the target area's air pollution data are used for calculation. The first monitoring point is located close to the pollution source, and the second monitoring point is located far from the pollution source. The difference between the air pollution concentrations at the first and second monitoring points is calculated to obtain the pollution concentration difference. The pollution concentration gradient is calculated by dividing the pollution concentration difference by the distance between the first and second monitoring points. The pollution diffusion coefficient is then calculated by multiplying the pollution concentration gradient by the wind speed.

[0072] Step S3: Construct a first composite feature based on the pollution diffusion coefficient, meteorological data, and air pollution data; construct a second composite feature based on the pollution diffusion coefficient, traffic congestion data, and meteorological data; and construct a third composite feature based on the soil physicochemical data and plant physiological data.

[0073] In practical applications, it is necessary to preprocess the collected meteorological data, air pollution data, soil physicochemical data, and plant physiological data. This includes cleaning operations such as handling missing and outlier values. Then, the cleaned meteorological data, air pollution data, soil physicochemical data, and plant physiological data are time-aligned and spatially aligned to ensure strict alignment in timestamps and spatial locations. Finally, the aligned meteorological data, air pollution data, soil physicochemical data, and plant physiological data are standardized to unify their dimensions and prevent features with large numerical ranges from dominating model training.

[0074] Specifically, the construction of the first composite feature based on the pollution diffusion coefficient, meteorological data, and air pollution data includes:

[0075] Wind speed and temperature sequences are extracted from meteorological data based on a preset sliding time window.

[0076] The pollution diffusion sequence is extracted from the pollution diffusion coefficient based on the sliding time window.

[0077] The pollution concentration sequence is extracted from the air pollution data based on the sliding time window.

[0078] The pollution diffusion sequence, wind speed sequence, temperature sequence, and pollution concentration sequence are combined to obtain the first composite feature.

[0079] In practical applications, it is necessary to pre-set the sliding time window of the collected data to a 24-hour sliding time window. Using this sliding time window, wind speed sequences are extracted from meteorological data, temperature sequences are extracted from meteorological data, and pollution concentration sequences are extracted from air pollution data. Then, the pollution diffusion sequence, wind speed sequence, temperature sequence, and pollution concentration sequence are sequentially concatenated to obtain the fusion feature matrix of meteorological data and air pollution data, i.e., the first composite feature.

[0080] Specifically, the construction of the second composite feature based on the pollution diffusion coefficient, traffic congestion data, and meteorological data includes:

[0081] The pollution diffusion sequence is extracted from the pollution diffusion coefficient based on a preset sliding time window;

[0082] Emission intensity data is calculated based on the traffic congestion data, and an emission intensity sequence is extracted from the emission intensity data based on the sliding time window.

[0083] Wind speed sequences are extracted from meteorological data based on the sliding time window.

[0084] The wind speed sequence, pollution diffusion sequence, and emission intensity sequence are combined to obtain the second composite feature.

[0085] In practical applications, the emission factors of pollutants are determined based on vehicle speeds in traffic congestion data, and the congestion duration is calculated based on the traffic congestion delay index. The emission factors of each pollutant are multiplied by the congestion duration to obtain the emission intensity data for each pollutant. Then, an emission intensity sequence is extracted from the emission intensity data using a 24-hour sliding time window. Finally, the wind speed sequence, pollution diffusion sequence, and emission intensity sequence are sequentially concatenated to obtain a fusion feature matrix of traffic data, meteorological data, and air pollution data, i.e., the second composite feature.

[0086] Specifically, the construction of the third composite feature based on the soil physicochemical data and plant physiological data includes:

[0087] The bioavailability of heavy metals was calculated based on the soil physicochemical data, and the plant stress resistance index and plant purification potential of the target plant were calculated based on the plant physiological data.

[0088] The ecological buffering capacity is obtained by calculating the geometric mean of the bioavailability, plant stress resistance index and plant purification potential.

[0089] The stress resistance index sequence is extracted from the plant stress resistance index according to the preset sliding time window, the buffer capacity sequence is extracted from the ecological buffer capacity according to the sliding time window, the bioavailability is replicated and extended to the time step dimension to obtain the bioavailability sequence, and the plant purification potential is replicated and extended to the time step dimension to obtain the purification potential sequence.

[0090] The bioavailability sequence, plant stress resistance index sequence, plant purification potential sequence, and buffering capacity sequence are combined to obtain the third composite feature.

[0091] In practical applications, the extractable heavy metal content in soil physicochemical data is divided by the total heavy metal content to obtain the extractable heavy metal ratio. This ratio is then multiplied by soil buffering capacity to obtain bioavailability. The POD activity in plant physiological data is multiplied by its corresponding contribution weight to obtain the POD contribution. This contribution is then added to SOD activity to obtain plant activity data. The MDA in plant physiological data is multiplied by its corresponding damage weight to obtain the MDA damage. This damage is then added to ROS to obtain plant damage data. The plant activity data is divided by the plant damage data to obtain the plant stress resistance index. The carbon-nitrogen ratio and leaf area index in plant physiological data are multiplied by the heavy metal enrichment coefficient in soil physicochemical data to obtain the plant purification potential. Then, a simple geometric mean is calculated for bioavailability, plant stress resistance index, and plant purification potential to enhance nonlinear interactions. Finally, the bioavailability sequence, plant stress resistance index sequence, plant purification potential sequence, and buffering capacity sequence are sequentially concatenated to obtain the fusion feature matrix of soil physicochemical data and plant physiological data, i.e., the third composite feature.

[0092] Step S4: Input the first composite feature, the second composite feature, and the third composite feature into the pre-trained pollution prediction model to obtain the pollution prediction result.

[0093] Specifically, the pollution prediction model is constructed based on a long short-term memory network, and the pollution prediction model includes:

[0094] The first channel constructed by the two-layer long short-term memory network is used to perform temporal analysis on the first composite feature of the input to obtain the first hidden state vector;

[0095] The second channel constructed by the single-layer long short-term memory network is used to perform feature concatenation and analysis on the input second composite feature and third composite feature to obtain the second hidden state vector.

[0096] The feature fusion layer is used to perform weighted fusion of the first hidden state vector and the second hidden state vector to obtain a weighted fusion vector.

[0097] The output layer is used to decode, temporally expand, and spatially reconstruct the weighted fusion vector to obtain the pollution prediction result.

[0098] In practical applications, the first channel serves as the primary channel, and the second channel as an auxiliary channel. The first channel dominates the prediction results, while the second channel corrects the predictions from the first channel. The first channel's two-layer long short-term memory network captures the physical diffusion process of pollutants in the atmosphere, with the first hidden state vector at the final time step representing the core diffusion patterns. The second channel's single-layer long short-term memory network learns the dynamic interaction between traffic emissions and ecosystem regulation (such as plant adsorption and soil degradation), with the second hidden state vector at the final time step representing the synergistic effect of emissions-ecological regulation. The feature fusion layer employs dynamic weighted fusion, concatenating the first and second hidden state vectors to obtain a concatenated state vector. An attention mechanism is used to generate a weight vector, assigning higher weights to key features such as wind speed and temperature. This results in a weighted fusion vector after passing through the feature fusion layer. Then, the pollution prediction result is obtained by decoding and temporal expansion of the weighted fusion vector. The first fully connected layer is used to decode the weighted fusion vector to extract high-order nonlinear relationships and obtain the decoded vector. The second fully connected layer is used to temporally expand the decoded vector to generate predictions for future time steps and obtain the temporal prediction vector. The temporal prediction vector is then spatially reconstructed and converted into a spatiotemporal grid structure to obtain the pollution prediction result.

[0099] Specifically, the first channel includes:

[0100] A first input layer is used to receive the first composite feature;

[0101] The first long short-term memory network layer is used to extract the hidden state vector of the first composite feature at each time step;

[0102] The second long short-term memory network layer is used to aggregate the hidden state vectors of all time steps across time steps to obtain the first hidden state vector of the final time step.

[0103] The first output layer is used to output the first hidden state vector.

[0104] In practical applications, the input dimension of the first composite feature is [sequence length, number of features]. For example, 24 hours × 4 features, where the 4 features are pollution diffusion coefficient, wind speed, temperature and pollution concentration. The first layer of the long short-term memory network captures local time patterns based on the first composite feature and obtains the hidden state vectors for each time step. The second layer of the long short-term memory network is used to integrate global temporal dependencies and obtain the first hidden state vector, focusing on the cumulative effect of historical sequences and providing a summary of diffusion trends for prediction.

[0105] The second channel includes:

[0106] The second input layer is used to receive the second composite feature and the third composite feature;

[0107] The feature splicing layer is used to splice the second composite feature and the third composite feature to obtain the spliced ​​feature;

[0108] The third long short-term memory network layer is used to aggregate the spliced ​​features across time steps to obtain the second hidden state vector of the final time step.

[0109] The output layer is used to output the second hidden state vector.

[0110] In practical applications, the input dimensions for both the second and third composite features are [sequence length, number of features]. These two composite features are then concatenated along the feature axis to form a unified sequence, resulting in a dimension of 24 hours × (3+4) features. The third long short-term memory network is used to learn the cross-cycle interactions between traffic emissions and ecological factors, yielding the second hidden state vector. Compared to meteorological diffusion, emissions and ecological processes change relatively slowly, requiring no deep time-series modeling. Therefore, the second channel employs a lightweight design using a single-layer long short-term memory network to adapt to the gradual changes in ecological processes.

[0111] Step S5: Control the plant maintenance system corresponding to the target plant based on the pollution prediction results.

[0112] In practical applications, a plant irrigation and maintenance system is set up in the target area. After obtaining the pollution prediction results, the plant irrigation and maintenance system can be activated in advance according to the pollution prediction results. For example, the plant irrigation and maintenance system can be activated 2 hours before severe pollution to irrigate the target plants. Or, the plant irrigation and maintenance system can be used to spray salicylic acid on the target plants 1 hour before severe pollution.

[0113] Specifically, to verify the effectiveness of this plant maintenance method, this embodiment takes Zhengzhou City as an example. Zhengzhou City was divided into lightly congested, moderately congested, and heavily congested areas according to the traffic congestion delay index. Taking the heavily congested area as an example, two sets of experiments were conducted, with the heavily congested area as the target area. In the first set of experiments, red photinia and large-leaved privet were planted in the heavily congested area. Red photinia was divided into a control group and an experimental group, and large-leaved privet was also divided into a control group and an experimental group, with red photinia and large-leaved privet as the target plants. If the pollution prediction result obtained using steps S1-S4 indicated severe pollution after 2 hours, the plant maintenance system was activated to irrigate the experimental groups of red photinia and large-leaved privet. The control groups of red photinia and large-leaved privet received no treatment. Figure 2 and Figure 3As shown, pre-irrigation treatment increased the activities of POD, SOD, and CAT, as well as the content of PRO, in Photinia serratifolia and Ligustrum lucidum, while reducing the content of MDA. Pre-irrigation treatment of Photinia serratifolia and Ligustrum lucidum 2 hours in advance under heavy pollution conditions helps to improve the target plants' tolerance to urban traffic pollution and alleviate the degree of oxidative damage.

[0114] The second experiment involved planting Pittosporum tobira and Platanus orientalis in a heavily congested area. Pittosporum tobira was divided into a control group and an experimental group, and Platanus orientalis was also divided into a control group and an experimental group. Both Pittosporum tobira and Platanus orientalis were used as target plants. If the pollution prediction result obtained using steps S1-S4 indicated severe pollution after one hour, the plant maintenance system was activated to spray the Pittosporum tobira and Platanus orientalis experimental groups with salicylic acid. The Pittosporum tobira and Platanus orientalis control groups received no treatment. Figure 4 and Figure 5 As shown, pre-treatment with salicylic acid increased the activities of POD, SOD, and CAT, as well as the content of PRO, in Pittosporum tobira and Platanus orientalis, while reducing the content of MDA. Under conditions of heavy pollution, pre-treatment with salicylic acid two hours in advance can help improve the target plants' tolerance to urban traffic pollution and alleviate the degree of oxidative damage.

[0115] The present invention provides a plant maintenance method based on traffic pollution. By combining traffic congestion data, meteorological data, soil physicochemical data, and plant physiological data, the method predicts the pollution results of the target area. This facilitates the early maintenance treatment of target plants in the target area based on the pollution prediction results, which helps to improve the target plants' tolerance to urban traffic pollution and alleviate the degree of oxidative damage.

[0116] like Figure 2 As shown, this embodiment of the invention also provides a plant maintenance system based on traffic pollution, comprising:

[0117] The data acquisition module 10 is used to acquire traffic congestion data, meteorological data, air pollution data, soil physicochemical data and plant physiological data of target plants in the target area, wherein the target plants are planted in the target area;

[0118] Calculation module 20 is used to calculate the pollution diffusion coefficient based on the meteorological data and air pollution data;

[0119] The feature construction module 30 is used to construct a first composite feature based on the pollution diffusion coefficient, meteorological data and air pollution data, construct a second composite feature based on the pollution diffusion coefficient, traffic congestion data and meteorological data, and construct a third composite feature based on the soil physicochemical data and plant physiological data.

[0120] The pollution prediction module 40 is used to input the first composite feature, the second composite feature and the third composite feature into a pre-trained pollution prediction model to obtain pollution prediction results.

[0121] The maintenance control module 50 is used to control the plant maintenance system corresponding to the target plant based on the pollution prediction results.

[0122] The various modules of the aforementioned plant maintenance system based on traffic pollution can be implemented entirely or partially through software, hardware, or a combination thereof. These modules and units can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0123] like Figure 3 As shown, an embodiment of the present invention discloses a computer device, including a memory and a processor, wherein the memory stores a computer program;

[0124] The computer device can be a server, and its internal structure diagram can be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and the database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the traffic pollution-based plant maintenance method described in the above embodiments.

[0125] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0126] This invention also discloses a computer-readable storage medium storing a computer program that causes a computer to execute the traffic pollution-based plant maintenance method described in the above embodiments.

[0127] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0128] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A method for plant care based on traffic pollution, characterized in that, The method comprises the following steps: acquiring traffic congestion data, meteorological data, air pollution data, soil physicochemical data, and plant physiological data of target plants planted in the target area; calculating a pollution diffusion coefficient based on the meteorological data and the air pollution data; constructing a first composite feature based on the pollution diffusion coefficient, the meteorological data, and the air pollution data, constructing a second composite feature based on the pollution diffusion coefficient, the traffic congestion data, and the meteorological data, and constructing a third composite feature based on the soil physicochemical data and the plant physiological data; inputting the first composite feature, the second composite feature, and the third composite feature into a pre-trained pollution prediction model to obtain a pollution prediction result; controlling a plant maintenance system corresponding to the target plants based on the pollution prediction result; the pollution prediction model is constructed based on a long short-term memory network, and the pollution prediction model comprises: a first channel constructed based on a double-layer long short-term memory network, configured to perform time series analysis on the input first composite feature to obtain a first hidden state vector; a second channel constructed based on a single-layer long short-term memory network, configured to perform feature splicing and analysis on the input second composite feature and third composite feature to obtain a second hidden state vector; a feature fusion layer, configured to perform weighted fusion on the first hidden state vector and the second hidden state vector to obtain a weighted fusion vector; an output layer, configured to perform decoding and time domain expansion on the weighted fusion vector to obtain the pollution prediction result.

2. The method for plant maintenance based on traffic pollution according to claim 1, wherein The pollution diffusion coefficient is calculated based on the meteorological data and the air pollution data, which comprises: calculating pollution gradient data based on the air pollution data; calculating the pollution diffusion coefficient based on the pollution gradient data and wind speed in the meteorological data.

3. The method for plant maintenance based on traffic pollution according to claim 1, wherein The first composite feature is constructed based on the pollution diffusion coefficient, the meteorological data, and the air pollution data, which comprises: extracting a wind speed sequence and a temperature sequence from the meteorological data based on a preset sliding time window; extracting a pollution diffusion sequence from the pollution diffusion coefficient based on the sliding time window; extracting a pollution concentration sequence from the air pollution data based on the sliding time window; combining the pollution diffusion sequence, the wind speed sequence, the temperature sequence, and the pollution concentration sequence to obtain the first composite feature.

4. The method for plant maintenance based on traffic pollution according to claim 1, wherein The second composite feature is constructed based on the pollution diffusion coefficient, the traffic congestion data, and the meteorological data, which comprises: extracting a pollution diffusion sequence from the pollution diffusion coefficient based on a preset sliding time window; calculating emission intensity data based on the traffic congestion data, and extracting an emission intensity sequence from the emission intensity data based on the sliding time window; extracting a wind speed sequence from the meteorological data based on the sliding time window; combining the wind speed sequence, the pollution diffusion sequence, and the emission intensity sequence to obtain the second composite feature.

5. The method for plant maintenance based on traffic pollution according to claim 1, wherein The third composite feature is constructed based on the soil physicochemical data and the plant physiological data, which comprises: calculating the bioavailability of heavy metals based on the soil physicochemical data, and calculating the plant stress resistance index and the plant purification potential of the target plants based on the plant physiological data; The biological effectiveness, the plant stress resistance index and the plant purification potential are geometrically averaged to obtain an ecological buffer capacity; According to a preset sliding time window, a stress resistance index sequence is extracted from the plant stress resistance index, a buffer capacity sequence is extracted from the ecological buffer capacity according to the sliding time window, the biological effectiveness is extended to a time step dimension to obtain an effectiveness sequence, and the plant purification potential is extended to the time step dimension to obtain a purification potential sequence; The biological effectiveness sequence, the plant stress resistance index sequence, the plant purification potential sequence and the buffer capacity sequence are combined to obtain a third composite feature.

6. The method for plant maintenance based on traffic pollution according to claim 1, wherein, The first channel comprises: a first input layer configured to receive the first composite feature; a first long short-term memory network layer configured to extract a hidden state vector of the first composite feature at each time step; a second long short-term memory network layer configured to aggregate the hidden state vectors of all time steps across time steps to obtain a first hidden state vector at a final time step; a first output layer configured to output the first hidden state vector; The second channel comprises: a second input layer configured to receive the second composite feature and the third composite feature; a feature splicing layer configured to splice the second composite feature and the third composite feature to obtain a spliced feature; a third long short-term memory network layer configured to aggregate the spliced feature across time steps to obtain a second hidden state vector at a final time step; an output layer configured to output the second hidden state vector.

7. A plant maintenance system based on traffic pollution, characterized by, It comprises: a data acquisition module configured to acquire traffic congestion data, meteorological data, air pollution data, soil physicochemical data and plant physiological data of target plants in a target region, wherein the target plants are planted in the target region; a calculation module configured to calculate a pollution diffusion coefficient according to the meteorological data and the air pollution data; a feature construction module configured to construct a first composite feature according to the pollution diffusion coefficient, the meteorological data and the air pollution data, to construct a second composite feature according to the pollution diffusion coefficient, the traffic congestion data and the meteorological data, and to construct a third composite feature according to the soil physicochemical data and the plant physiological data; a pollution prediction module configured to input the first composite feature, the second composite feature and the third composite feature into a pre-trained pollution prediction model to obtain a pollution prediction result; a maintenance control module configured to control a plant maintenance system corresponding to the target plants according to the pollution prediction result; The pollution prediction model is constructed based on a long short-term memory network, and the pollution prediction model comprises: a first channel constructed by a double-layer long short-term memory network, configured to perform time series analysis on the input first composite feature to obtain a first hidden state vector; a second channel constructed by a single-layer long short-term memory network, configured to perform feature splicing and analysis on the input second composite feature and third composite feature to obtain a second hidden state vector; a feature fusion layer configured to perform weighted fusion on the first hidden state vector and the second hidden state vector to obtain a weighted fusion vector; an output layer configured to decode and time-domain expand the weighted fusion vector to obtain the pollution prediction result.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1-6.

Images