Method and system for observing greenhouse gas source of underway city and application

Abstract

The invention discloses an observation method and system for a navigation city greenhouse gas source and application, belongs to the field of environment monitoring, and is used for solving the technical problem that at present, effective and targeted monitoring on the city greenhouse gas source is lacked, and the method comprises the steps that an observation route for monitoring the greenhouse gas source in a navigation city is planned; constructing a navigation city greenhouse gas source distribution prediction model and training, carrying out current navigation monitoring to obtain current navigation monitoring data, and calculating by using the navigation city greenhouse gas source distribution prediction model to obtain next city greenhouse gas source prediction distribution; performing targeted underway monitoring planning and monitoring; according to the method, the observation route is planned, multiple times of underway monitoring are carried out according to the time sequence, accurate underway monitoring data are obtained, the data are fully mined, the prediction model is constructed and trained, the model is used for calculation, the next prediction distribution is obtained, the model is used for guiding subsequent targeted underway monitoring, and the monitoring efficiency and effect of the urban greenhouse gas source are improved.

Description

Technical Field

[0001] This application belongs to the field of environmental monitoring, specifically the observation methods, systems and applications for mobile urban greenhouse gas sources. Background Technology

[0002] With the increasing prominence of global climate change, urban areas, as major sources of greenhouse gas emissions (such as CO2 and CH4), have made emission monitoring and source tracing a core task of environmental governance. Currently, urban greenhouse gas monitoring mainly relies on fixed-site observations, satellite remote sensing, and preliminary mobile observations. Fixed-site observations can achieve long-term continuous monitoring, but the coverage is limited and it is difficult to reflect the spatial heterogeneity of greenhouse gases within the city (such as the concentration difference between industrial areas and residential areas); satellite remote sensing can obtain information on large-scale distribution, but the spatial resolution is low (usually at the kilometer level) and is easily affected by clouds and buildings, making it difficult to capture small-scale emission sources within the city (such as gas stations, landfills and other point sources).

[0003] Currently, there is a lack of effective and targeted monitoring of urban greenhouse gas sources. Summary of the Invention

[0004] This application aims to address at least one of the technical problems existing in the prior art; to this end, this application proposes a mobile observation method, system and application for urban greenhouse gas sources, in order to solve the current lack of effective and targeted monitoring of urban greenhouse gas sources.

[0005] To achieve the above objectives, the first aspect of this application provides a method for observing mobile urban greenhouse gas sources, including, Step 1: Obtain the distribution information of historical urban greenhouse gas sources, and based on the distribution information of historical urban greenhouse gas sources, plan the observation route for mobile monitoring of urban greenhouse gas sources. Step 2: Based on the planned observation route for mobile urban monitoring of greenhouse gas sources, multiple mobile monitoring operations are conducted in a time sequence using off-axis integrating cavity spectroscopy technology to obtain mobile monitoring data, which is then collected into a mobile urban monitoring dataset. A training sample set is constructed based on the mobile urban monitoring dataset, which includes several mobile monitoring data based on time steps. Step 3: Construct and train a mobile urban greenhouse gas source distribution prediction model. The input of the mobile urban greenhouse gas source distribution prediction model is the mobile monitoring data of the first N times, and the output is the predicted distribution of urban greenhouse gas sources for N+1 times. Step 4: Conduct the current mobile monitoring to obtain the current mobile monitoring data, and use the mobile urban greenhouse gas source distribution prediction model to calculate the predicted distribution of urban greenhouse gas sources for the next time. Step 5: Based on the predicted distribution of urban greenhouse gas sources, conduct targeted mobile monitoring planning and monitoring.

[0006] Another aspect of this application discloses a mobile observation system for urban greenhouse gas sources, including a mobile monitoring unit for performing mobile observations; The cloud server is used to store the data uploaded by the mobile monitoring unit; The observation route planning module is used to plan the observation route for mobile city monitoring of greenhouse gas sources based on the distribution information of the historical urban greenhouse gas sources. The prediction model building and training module is used to build and train a mobile urban greenhouse gas source distribution prediction model. And a targeted mobile monitoring planning unit, used to conduct targeted mobile monitoring planning based on the next urban greenhouse gas source prediction distribution and off-axis integrating cavity spectroscopy technology.

[0007] Another aspect of this application discloses the application of the above-mentioned mobile urban greenhouse gas source observation system in urban greenhouse gas source observation.

[0008] Compared with existing technologies, the beneficial effects of this application are as follows: Based on the distribution information of historical urban greenhouse gas sources, this application plans the observation route for mobile urban monitoring of greenhouse gas sources, which facilitates rapid and accurate targeted monitoring. At the same time, this application conducts multiple mobile monitoring operations in a time sequence based on off-axis integrating cavity spectroscopy technology to obtain accurate mobile monitoring data. Furthermore, it fully mines the mobile monitoring data to construct and train a mobile urban greenhouse gas source distribution prediction model. Using this model, the predicted distribution of urban greenhouse gas sources for the next time is obtained, which can guide subsequent targeted mobile monitoring and improve the efficiency and effectiveness of urban greenhouse gas source monitoring. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating the observation method for mobile urban greenhouse gas sources in this application.

[0011] Figure 2 This is a schematic diagram of the observation route for mobile urban monitoring of greenhouse gas sources in this embodiment of the application.

[0012] Figure 3This is a structural block diagram of the mobile urban greenhouse gas source observation system in the embodiments of this application.

[0013] Wherein: 10, Urban greenhouse gas source distribution map; 11, Point source; 12, Line source; L, Observation route; 20, Observation system; 21, Mobile monitoring unit; 211, Mobile platform; 212, Off-axis integrating cavity spectrometer; 213, High-precision vehicle-mounted base station differential GPS; 214, Three-dimensional ultrasonic anemometer; 215, Operation terminal; 22, Observation route planning module; 23, Prediction model construction and training module; 24, Cloud server; 25, Targeted mobile monitoring planning unit. Detailed Implementation

[0014] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0015] Please see Figure 1-3 The first aspect of this application provides a method for observing mobile urban greenhouse gas sources, including: Step 1: Obtain the distribution information of historical urban greenhouse gas sources, and based on the distribution information of historical urban greenhouse gas sources, plan the observation route for mobile monitoring of urban greenhouse gas sources. In an exemplary embodiment, the method for extracting the distribution information of historical urban greenhouse gas sources in step one, and planning the observation route for mobile urban monitoring of greenhouse gas sources based on the distribution information of historical urban greenhouse gas sources, includes: Historical distribution information of urban greenhouse gas sources is extracted and analyzed to obtain point sources and line sources corresponding to urban greenhouse gas sources. The locations of point sources include industrial areas, gas stations, biochemical pools, and landfills; the locations of line sources include road ring roads. For example, a city greenhouse gas source distribution map 10 is constructed based on the point sources and line sources, and an observation route L for mobile city monitoring of greenhouse gas sources is planned based on the point sources 11 and line sources 12. Step 2: Based on the planned observation route L of the mobile urban monitoring greenhouse gas source, multiple mobile monitoring operations are carried out in time sequence using off-axis integrating cavity spectroscopy technology to obtain mobile monitoring data, which are then collected into a mobile urban monitoring dataset. A training sample set is constructed based on the mobile urban monitoring dataset, which includes several mobile monitoring data based on time step sequence. In an exemplary embodiment, the method for performing multiple time-series mobile monitoring based on off-axis integrating cavity spectroscopy includes, The mobile monitoring unit 21 performs multiple mobile monitoring operations along the planned observation route L in a time sequence, with the time between two adjacent mobile monitoring operations being a time step; in an exemplary embodiment, the time step is 3 hours, and in some embodiments, the time step can be 8 hours or 12 hours.

[0016] The mobile monitoring unit includes at least one vehicle, which is configured as a mobile platform 211. The monitoring equipment mounted on the mobile platform 211 includes an off-axis integrating cavity spectrometer 212, the air inlet of which is located on the top of the mobile platform 211. The spectrometer is used to analyze the concentration of urban greenhouse gas sources and upload the data to a cloud server. The urban greenhouse gas sources include CO2 and CH4. The high-precision vehicle-mounted base station differential GPS213, located at the top of the mobile platform, is used to acquire latitude and longitude coordinates and upload the data to the cloud server. The 214 three-dimensional ultrasonic anemometer, located on top of the mobile platform, is used to acquire wind speed and direction data and upload the data to the cloud server. And an operation terminal 215, which is a laptop computer, used to retrieve data from the cloud server in real time or at regular intervals, and to judge and record the operating status of the unit on-site or remotely.

[0017] Step 3: Construct and train a mobile urban greenhouse gas source distribution prediction model. The input of the mobile urban greenhouse gas source distribution prediction model is the mobile monitoring data of the first N times, and the output is the predicted distribution of urban greenhouse gas sources for N+1 times. In one exemplary embodiment, N is 3; in some embodiments, N can be 5. In an exemplary embodiment, the mobile urban greenhouse gas source distribution prediction model is a spatiotemporal fusion deep learning model. Specifically, considering that the mobile monitoring data has the dual characteristics of continuous correlation in time steps and dynamic changes in spatial location, and that subsequent distributions need to be derived based on the previous N monitoring data, which may contain multi-dimensional information such as CO2 and CH4 concentrations, latitude and longitude, wind speed and direction, the model needs to have both time-series dependency capture capability and spatial feature fusion capability. Therefore, a hybrid architecture of time-series prediction network and spatial feature extraction module is selected as the basic framework. In an exemplary embodiment, the mobile urban greenhouse gas source distribution prediction model includes a data preprocessing layer, which is used to standardize and structure the data from the previous N mobile monitoring sessions, and reorganize them into a multi-dimensional temporal spatial matrix according to the time step sequence. The temporal feature extraction layer is used to capture the temporal dependencies of the previous N mobile monitoring data through a bidirectional long short-term memory network. The spatial feature fusion layer is used to extract and fuse spatial features by combining convolutional neural network kernels and attention mechanisms. The prediction output layer is used to concatenate temporal and spatial features, and then map them into a spatial grid feature matrix of N+1 monitoring through a fully connected layer. After inverse normalization, it is transformed into concentration distribution data under actual geographic coordinates, and outputs the N+1 predicted distribution of urban greenhouse gas sources. The N+1 predicted distribution of urban greenhouse gas sources includes spatial grids, corresponding concentration intervals, and time periods.

[0018] Specifically, the output features of the temporal feature extraction layer and the spatial feature fusion layer are concatenated and mapped to a spatial grid feature matrix of N+1 monitoring through a fully connected layer. Then, combined with inverse normalization processing, the predicted feature matrix is ​​transformed into greenhouse gas concentration distribution data under actual geographic coordinates. Finally, the N+1 predicted distribution of urban greenhouse gas sources is output with spatial grid, corresponding concentration interval and time period as the main components. In some embodiments, the N+1th urban greenhouse gas source prediction distribution may also include a corresponding confidence level; the confidence level is calculated by the error between the model prediction value and the validation set in the training sample set, and is used to evaluate the reliability of the prediction results. In an exemplary embodiment, the method for standardizing and structuring the data from the first N mobile monitoring operations includes standardizing the greenhouse gas concentration data, converting latitude and longitude into Cartesian coordinates, and dividing the data into spatial grids; and decomposing wind speed and direction into east-west and north-south components.

[0019] Specifically, the first N mobile monitoring data in the training sample set are standardized: for greenhouse gas concentration data, standardization can be used to eliminate the magnitude interference caused by the difference in accuracy of different monitoring equipment; for latitude and longitude data, Gaussian projection transformation is performed to convert geographical coordinates into plane rectangular coordinates, which facilitates subsequent spatial grid division; for wind speed and direction data, vector decomposition is used to convert them into east-west and north-south wind speed components to avoid nonlinear errors caused by direct calculation of direction angles.

[0020] Meanwhile, according to the time step order, the first N data are divided into one time slice for each monitoring data, and each slice is reorganized in the form of multi-dimensional features within the spatial grid to form the temporal spatial matrix of the model input. The dimension can be N×M×K, where M is the number of spatial grids divided by the navigation route and K is the feature dimension of a single monitoring. In an exemplary embodiment, let the time-series spatial matrix formed by the preprocessing of the first N mobile monitoring data be the input variable X, and let the mathematical expression of X be X=[X1,X2,...,Xi,...,XN]ϵR N×M×K ; R N×M×KIt is a three-dimensional real matrix space that includes time, space, and features; N is the number of time steps input (the first N mobile monitoring cycles); Xi, the spatial feature matrix of the i-th mobile monitoring, XiϵR M×K , i=1,2,...,N; M represents the number of spatial grids divided by the navigation route; K is the feature dimension of a single monitoring. In an exemplary embodiment, K=6, including CO2 concentration and CH4 concentration, plane coordinate x, plane coordinate y, east-west wind speed component and north-south wind speed component. Let Y be the N+1th predicted distribution of urban greenhouse gas sources output by the mobile urban greenhouse gas source distribution prediction model, and let Y be mathematically expressed as Y=[Y 1,1 ,Y 1,2 ,Y 2,1 ,Y 2,2 ,...,Y m,c ,...,X M,2 ]ϵR M×2 ; M is the number of spatial grids divided by the navigation route, which is consistent with the input. c=1,2 correspond to CO2 and CH4 respectively; Y m,c This represents the predicted concentration of the c-th greenhouse gas in the m-th spatial grid. R M×2 Represents a two-dimensional real matrix space containing the number of spatial grids M for the navigation route division and two characteristics: CO2 and CH4; In an exemplary embodiment, the method for capturing the temporal dependencies of the previous N mobile monitoring data through a bidirectional long short-term memory network further includes identifying the changing patterns of greenhouse gas concentrations over time from the forward and reverse time series, and suppressing overfitting by using a dropout layer.

[0021] Specifically, a bidirectional long short-term memory network (Bi-LSTM) can be used as the core component. This network can simultaneously capture the temporal dependencies of monitoring data from both the forward time series (from the 1st to the Nth time) and the backward time series (from the Nth to the 1st time). For the hidden layer settings of the Bi-LSTM, combined with the time step of urban mobile monitoring, for example, the time step can be set to 3 hours / time, with 5-8 monitoring sessions per day. The number of neurons in the hidden layer can be set to 256, and the size of the time step should be consistent with the time step of the previous N monitoring sessions. Overfitting is suppressed by using dropout layers, where the dropout probability can be set to 0.2. This ensures that the model has the ability to generalize to temporal changes under different seasons and weather conditions. In an exemplary embodiment, the method for extracting and fusing spatial features by means of convolutional neural network kernel combination and attention mechanism includes enhancing the correlation of feature channels by using a first convolutional kernel, capturing the interaction of adjacent grids by a second convolutional kernel, and assigning higher weights to the grids where high emission point sources are located; the first convolutional kernel is a 1×1 convolutional kernel; and the second convolutional kernel is a 3×3 convolutional kernel.

[0022] Specifically, a combination of 1×1 and 3×3 convolutional kernels from a convolutional neural network (CNN) is introduced to extract spatial features from the temporal feature matrix output by the Bi-LSTM: the 1×1 convolutional kernel is used to compress feature dimensions and enhance the correlation between channels (such as the coupling relationship between CO2 concentration and wind speed within a spatial grid); the 3×3 convolutional kernel is used to capture the feature interactions between adjacent spatial grids (such as the diffusion effect of vehicle exhaust emissions from a road ring on the concentration of surrounding industrial areas). At the same time, an attention mechanism is embedded to dynamically allocate feature weights to different spatial grids. For example, higher attention weights are given to the grids where point sources with excessive concentrations in historical data (such as gas stations and biochemical pools) are located, ensuring that the model prioritizes the spatial features of high-emission areas. Step 4: Conduct the current mobile monitoring to obtain the current mobile monitoring data, and use the mobile urban greenhouse gas source distribution prediction model to calculate the predicted distribution of urban greenhouse gas sources for the next time. In an exemplary embodiment, the mobile monitoring unit 21 is used to conduct the current mobile monitoring along the planned observation route L to obtain the current mobile monitoring data. The mobile monitoring data is then combined with the mobile monitoring data from the previous N-1 mobile monitoring sessions, for a total of N mobile monitoring data, as input and calculated using the mobile urban greenhouse gas source distribution prediction model to obtain the predicted distribution of urban greenhouse gas sources for the next time. Step 5: Based on the next predicted distribution of urban greenhouse gas sources, conduct targeted mobile monitoring planning and monitoring. In an exemplary embodiment, the method for conducting targeted mobile monitoring planning and monitoring based on the next predicted distribution of urban greenhouse gas sources includes: Based on the predicted distribution of urban greenhouse gas sources, key locations and corresponding time periods exceeding preset thresholds are extracted, and a key mobile monitoring scheme is planned and generated. Mobile monitoring is then carried out on the key locations and corresponding time periods exceeding preset thresholds according to the key mobile monitoring scheme.

[0023] In an exemplary embodiment, the preset threshold can be set in combination with greenhouse gas type, urban functional zoning characteristics, environmental management objectives, and historical monitoring data patterns.

[0024] This application discloses, by way of example, a mobile observation system 20 for urban greenhouse gas sources, including a mobile monitoring unit 21 for performing mobile observations; The observation route planning module 22 is used to plan the observation route for mobile city monitoring of greenhouse gas sources based on the distribution information of the historical urban greenhouse gas sources. Prediction model building and training module 23 is used to build and train a mobile urban greenhouse gas source distribution prediction model. Cloud server 24 is used to store the data uploaded by the mobile monitoring unit; And a targeted mobile monitoring planning unit 25, used to conduct targeted mobile monitoring planning based on off-axis integrating cavity spectroscopy technology, according to the next predicted distribution of urban greenhouse gas sources.

[0025] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A method for observing urban greenhouse gas sources by mobile surveying, characterized in that, include, Step 1: Obtain the distribution information of historical urban greenhouse gas sources, and based on the distribution information of historical urban greenhouse gas sources, plan the observation route for mobile monitoring of urban greenhouse gas sources. Step 2: Based on the planned observation route for mobile urban monitoring of greenhouse gas sources, multiple mobile monitoring operations are conducted in a time sequence using off-axis integrating cavity spectroscopy technology to obtain mobile monitoring data, which is then collected into a mobile urban monitoring dataset. A training sample set is constructed based on the mobile urban monitoring dataset, which includes several mobile monitoring data based on time steps. Step 3: Construct and train a mobile urban greenhouse gas source distribution prediction model. The input of the mobile urban greenhouse gas source distribution prediction model is the mobile monitoring data of the first N times, and the output is the predicted distribution of urban greenhouse gas sources for N+1 times. Step 4: Conduct the current mobile monitoring to obtain the current mobile monitoring data, and use the mobile urban greenhouse gas source distribution prediction model to calculate the predicted distribution of urban greenhouse gas sources for the next time. Step 5: Based on the predicted distribution of urban greenhouse gas sources, conduct targeted mobile monitoring planning and monitoring.

2. The method for observing urban greenhouse gas sources by mobile monitoring according to claim 1, characterized in that, The method for performing multiple time-series mobile monitoring based on off-axis integrating cavity spectroscopy includes... The mobile monitoring unit conducts multiple mobile monitoring operations along the planned observation route in a time sequence, with the time between two adjacent mobile monitoring operations constituting one time step. The mobile monitoring unit includes at least one vehicle, which is configured as a mobile platform. The monitoring equipment mounted on the mobile platform includes an off-axis integrating cavity spectrometer, the air inlet of which is located at the top of the mobile platform. The spectrometer is used to analyze the concentration of urban greenhouse gas sources and upload the data to a cloud server. The urban greenhouse gas sources include CO2 and CH4. The high-precision vehicle-mounted base station differential GPS, located at the top of the mobile platform, is used to acquire latitude and longitude coordinates and upload the data to the cloud server; The three-dimensional ultrasonic anemometer, located on top of the mobile platform, is used to acquire wind speed and direction data and upload the data to the cloud server. And an operating terminal, which is a laptop computer, used to retrieve data from the cloud server in real time or at regular intervals, and to judge and record the operating status of the unit on-site or remotely.

3. The method for observing urban greenhouse gas sources by mobile monitoring according to claim 2, characterized in that, Step one involves extracting distribution information of historical urban greenhouse gas sources. Based on this information, the method for planning observation routes for mobile urban monitoring of greenhouse gas sources includes... Historical distribution information of urban greenhouse gas sources is extracted and analyzed to obtain point sources and line sources corresponding to urban greenhouse gas sources. The locations of point sources include industrial areas, gas stations, biochemical pools, and landfills; the locations of line sources include road loops. Based on the point sources and line sources, a mobile monitoring route for urban greenhouse gas sources is planned.

4. The method for observing urban greenhouse gas sources by mobile monitoring according to claim 3, characterized in that, The mobile monitoring model for predicting the distribution of greenhouse gas sources in cities is a spatiotemporal fusion deep learning model. The mobile monitoring model for predicting the distribution of greenhouse gas sources in cities includes a data preprocessing layer, which is used to standardize and structure the data from the previous N mobile monitoring sessions, and reorganize them into a multi-dimensional temporal spatial matrix according to the time step sequence. The temporal feature extraction layer is used to capture the temporal dependencies of the previous N mobile monitoring data through a bidirectional long short-term memory network. The spatial feature fusion layer is used to extract and fuse spatial features by combining convolutional neural network kernels and attention mechanisms. The prediction output layer is used to concatenate temporal and spatial features, and then map them into a spatial grid feature matrix of N+1 monitoring through a fully connected layer. After inverse normalization, it is transformed into concentration distribution data under actual geographic coordinates, and outputs the N+1 predicted distribution of urban greenhouse gas sources. The N+1 predicted distribution of urban greenhouse gas sources includes spatial grids, corresponding concentration intervals, and time periods.

5. The method for observing mobile urban greenhouse gas sources according to claim 4, characterized in that, The methods for standardizing and structuring the data from the previous N mobile monitoring operations include standardizing greenhouse gas concentration data, converting latitude and longitude into Cartesian coordinates, and dividing the data into spatial grids; and decomposing wind speed and direction into east-west and north-south components.

6. The method for observing mobile urban greenhouse gas sources according to claim 4, characterized in that, The method of capturing the temporal dependencies of the previous N mobile monitoring data through bidirectional long short-term memory networks also includes identifying the changes in greenhouse gas concentration over time from forward and reverse time series, and suppressing overfitting by using a dropout layer.

7. The method for observing urban greenhouse gas sources by mobile monitoring according to claim 4, characterized in that, Methods for extracting and fusing spatial features by combining convolutional neural network kernels and attention mechanisms include enhancing the correlation of feature channels through the first convolutional kernel, capturing the interaction between adjacent grids through the second convolutional kernel, and assigning higher weights to the grids where high emission point sources are located. The first convolution kernel is a 1×1 convolution kernel; the second convolution kernel is a 3×3 convolution kernel.

8. The method for observing urban greenhouse gas sources by mobile monitoring according to claim 1, characterized in that, Based on the predicted distribution of urban greenhouse gas sources, the methods for planning and monitoring targeted mobile monitoring include: Based on the predicted distribution of urban greenhouse gas sources, key locations and corresponding time periods exceeding preset thresholds are extracted, and a key mobile monitoring scheme is planned and generated. Mobile monitoring is then carried out on the key locations and corresponding time periods exceeding preset thresholds according to the key mobile monitoring scheme.

9. A mobile observation system for urban greenhouse gas sources using the method described in any one of claims 1-8, characterized in that, Includes a mobile monitoring unit for performing mobile observations; The cloud server is used to store the data uploaded by the mobile monitoring unit; The observation route planning module is used to plan the observation route for mobile city monitoring of greenhouse gas sources based on the distribution information of the historical urban greenhouse gas sources. The prediction model building and training module is used to build and train a mobile urban greenhouse gas source distribution prediction model. And a targeted mobile monitoring planning unit, used to conduct targeted mobile monitoring planning based on the next urban greenhouse gas source prediction distribution and off-axis integrating cavity spectroscopy technology.

10. The application of the mobile urban greenhouse gas source observation system according to claim 9 in urban greenhouse gas source observation.

Images