Multi-source data fusion method and device for road carbon emission monitoring

Abstract

The invention discloses a multi-source data fusion method and device for road carbon emission monitoring. The method comprises the following steps: collecting carbon emission multi-source data based on monitoring equipment on an underway monitoring vehicle; carrying out preprocessing on the carbon emission multi-source data; constructing an environment model according to the preprocessed carbon emission multi-source data, acquiring the acquisition time of the carbon emission multi-source data, acquiring spatial information according to the constructed environment model, and performing time matching and spatial matching on the carbon emission multi-source data based on the acquisition time and the spatial information; obtaining a carbon emission space-time distribution result based on the carbon emission multi-source data of time matching and space matching; through data fusion of the greenhouse gas concentration sensor, the GPS, the meteorological sensor, the panoramic camera, the laser radar measurement equipment and the like, high-precision space-time matching is realized, the difference of a single data source in spatial distribution and time points is effectively eliminated, and the accuracy of urban road traffic greenhouse gas monitoring is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and particularly to a multi-source data fusion method and device for road carbon emission monitoring. Background Art

[0002] With the increasingly severe global climate change problem, the monitoring and management of urban carbon emissions have become an important task for reducing greenhouse gas emissions and promoting sustainable development. Especially in the field of urban transportation, transportation is one of the main sources of carbon emissions. Therefore, accurate and real-time monitoring of urban road carbon emissions is crucial.

[0003] Currently, most existing road traffic greenhouse gas monitoring methods use fixed monitoring stations, that is, by deploying fixed monitoring points in specific road sections or urban areas, collecting data such as the CO2 concentration in the atmosphere, and combining traffic flow and traffic conditions for emission estimation. This method can provide certain carbon concentration monitoring data. However, its disadvantages are also very obvious. For example, fixed monitoring stations can only monitor the atmospheric CO2 concentration level at the set positions and cannot cover all roads comprehensively, resulting in the possible omission of data for some high-traffic sections or specific sections.

[0004] In addition, carbon emissions can also be calculated theoretically. For example, patent text CN117727185A discloses a carbon emission monitoring method based on vehicle flow data, including obtaining the entry image and exit image of a vehicle on a specified road, both of which carry the corresponding image generation time; obtaining the vehicle model and total driving time based on the entry image, exit image, and the image generation time carried by both, and determining the air-fuel ratio and its associated pollutant emission concentration volume ratio, pollutant volume mass conversion coefficient, and pollutant carbon content according to the vehicle model, and further combining the total driving time of the vehicle and the pre-stored relevant lengths of flat roads, uphill roads, and downhill roads in the specified road to obtain the pollutant emissions of the vehicle. However, this method relies on experience and has low accuracy.

[0005] With the development of sensor technology and mobile computing technology, in-vehicle mobile monitoring systems have gradually become a new means of carbon emission monitoring. By installing greenhouse gas monitoring equipment on vehicles, dynamic real-time monitoring of greenhouse gases such as CO2 and atmospheric pollutant concentrations on roads can be achieved. This method can provide a wider coverage range than fixed monitoring stations, obtain real carbon emission data, and can be flexibly monitored according to real-time road conditions. For example, conducting bypass monitoring for heavily polluting enterprises, etc. However, due to different types of data having different collection frequencies and spatial distributions, how to effectively match these data from different sensors in space and time and extract accurate spatio-temporal carbon concentration information from them remains an urgent problem to be solved. Summary of the Invention

[0006] The present invention provides a multi-source data fusion method and device for road carbon emission monitoring, which can obtain accurate spatio-temporal carbon concentration information.

[0007] A multi-source data fusion method for road carbon emission monitoring includes: Collecting multi-source carbon emission data based on monitoring devices on a mobile monitoring vehicle; Preprocessing the multi-source carbon emission data; Constructing an environmental model according to the preprocessed multi-source carbon emission data, obtaining the collection time of the multi-source carbon emission data, obtaining spatial information according to the constructed environmental model, and performing time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information; Obtaining a carbon emission spatio-temporal distribution result based on the multi-source carbon emission data that has undergone time matching and spatial matching.

[0008] Further, the multi-source carbon emission data includes greenhouse gas data, mobile vehicle position data, meteorological data, panoramic environmental image data, and environmental point cloud data.

[0009] Further, preprocessing the multi-source carbon emission data includes: Calculating the standard deviation and mean for the same type of data in the multi-source carbon emission data; Calculating the Z-score for each data point according to the standard deviation and mean, judging whether the data point is an abnormal data point according to the Z-score, and deleting the abnormal data points; Interpolating and filling in the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method.

[0010] Further, interpolating and filling in the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method includes: Sorting each type of data belonging to the time series according to the collection time; Determining the missing points in the data belonging to the time series according to the collection frequency of each type of data; Calculating the value of the missing point according to the value of the previous data point and the value of the subsequent data point of the missing point.

[0011] Further, constructing an environmental model according to the preprocessed multi-source carbon emission data, obtaining the collection time of the multi-source carbon emission data, obtaining spatial information according to the constructed environmental model, and performing time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information includes: Identifying the vehicles in the panoramic environmental image data by using a neural network model to obtain the vehicle types and quantities; Construct an environmental three-dimensional point cloud model based on the environmental point cloud data, and extract road vector data from the environmental three-dimensional point cloud model to form a first data set, where the road vector data includes the road point cloud acquisition time and the road position coordinates; Resample the greenhouse gas data and the meteorological data according to a preset time window to obtain greenhouse gas resampled data and meteorological resampled data under the time window; Set time stamps and position coordinates for the vehicle type and quantity, the greenhouse gas resampled data, and the meteorological resampled data according to the acquisition time of the panoramic environmental image data, the greenhouse gas resampled data, and the meteorological resampled data and the corresponding vehicle position data at the acquisition time, and form a second data set; Match the first data set and the second data set based on the nearest neighbor interpolation method or the inverse distance weighted method.

[0012] Further, matching the first data set and the second data set based on the nearest neighbor interpolation method includes: Select a data point to be matched from the first data set as a target point, and calculate the Euclidean distance and time difference between the target point and each data point in the second data set; Select an optimal data point from the data points in the second data set where the Euclidean distance and time difference meet the preset conditions to associate with the target point, and obtain a matching result.

[0013] Further, matching the first data set and the second data set based on the inverse distance weighted method includes: Select a data point to be matched from the first data set as a target point, and calculate the Euclidean distance and time difference between the target point and each data point in the second data set; Set an initial search radius, and for the target point, count the number of data points in the second data set within the initial search radius where the Euclidean distance and time difference meet the preset conditions, and calculate the local point density; Adjust the search radius according to the local point density preset density; Select the data points in the second data set that meet the preset conditions within the adjusted search radius, and perform distance weighted calculation according to the attribute values of the selected data points and their Euclidean distances from the target point to obtain an attribute estimation value; Associate the attribute estimation value with the target point to obtain a matching result.

[0014] Further, the preset conditions are that the Euclidean distance is less than a preset distance and the time difference is less than a preset duration.

[0015] A multi-source data fusion device for road carbon emission monitoring includes: A collection module for collecting multi-source carbon emission data based on monitoring devices on a vehicle for on-road monitoring; A preprocessing module for preprocessing the multi-source carbon emission data; A matching module for constructing an environmental model based on the preprocessed multi-source carbon emission data, obtaining the collection time of the multi-source carbon emission data, obtaining spatial information according to the constructed environmental model, and performing time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information; A result processing module for obtaining a spatio-temporal distribution result of carbon emissions based on the multi-source carbon emission data that has been time-matched and space-matched.

[0016] Furthermore, the multi-source carbon emission data includes greenhouse gas data, vehicle position data for on-road monitoring, meteorological data, panoramic environmental image data, and environmental point cloud data.

[0017] Furthermore, the preprocessing module preprocesses the multi-source carbon emission data, including: Calculating the standard deviation and mean for the same type of data in the multi-source carbon emission data; Calculating the Z-score for each data point according to the standard deviation and mean, determining whether the data point is an abnormal data point based on the Z-score, and deleting the abnormal data points; Performing interpolation filling on the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method.

[0018] Furthermore, the preprocessing module performs interpolation filling on the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method, including: Sorting each type of data belonging to the time series according to the collection time respectively; Determining the missing points in the data belonging to the time series according to the collection frequency of each type of data; Calculating the value of the missing point according to the value of the previous data point and the value of the subsequent data point of the missing point.

[0019] Furthermore, the matching module constructs an environmental model according to the preprocessed multi-source carbon emission data, obtains the collection time of the multi-source carbon emission data, obtains spatial information according to the constructed environmental model, and performs time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information, including: Identifying the vehicles in the panoramic environmental image data using a neural network model to obtain the vehicle types and quantities; Construct an environmental three-dimensional point cloud model based on the environmental point cloud data, and extract road vector data from the environmental three-dimensional point cloud model to form a first data set, where the road vector data includes the road point cloud acquisition time and the road position coordinates; Resample the greenhouse gas data and meteorological data according to a preset time window to obtain greenhouse gas resampled data and meteorological resampled data under the time window; Set timestamps and position coordinates for vehicle types and quantities, greenhouse gas resampled data, and meteorological resampled data according to the acquisition time of the panoramic environmental image data, greenhouse gas resampled data, and meteorological resampled data and the corresponding driving vehicle position data at that acquisition time, and form a second data set; Match the first data set and the second data set based on the nearest neighbor interpolation method or the inverse distance weighted method.

[0020] Further, the matching module matches the first data set and the second data set based on the nearest neighbor interpolation method, including: Select the data points to be matched from the first data set as target points, and calculate the Euclidean distance and time difference between the target points and each data point in the second data set; Select the optimal data point from the data points in the second data set where the Euclidean distance and time difference meet the preset conditions to associate with the target point, and obtain the matching result.

[0021] Further, the matching module matches the first data set and the second data set based on the inverse distance weighted method, including: Select the data points to be matched from the first data set as target points, and calculate the Euclidean distance and time difference between the target points and each data point in the second data set; Set an initial search radius, for the target point, count the number of data points in the second data set within the initial search radius where the Euclidean distance and time difference meet the preset conditions, and calculate the local point density; Adjust the search radius according to the local point density preset density; Select the data points in the second data set that meet the preset conditions within the adjusted search radius, and perform distance weighted calculation according to the attribute values of the selected data points and their Euclidean distances from the target point to obtain an attribute estimation value; Associate the attribute estimation value with the target point to obtain the matching result.

[0022] Further, the preset conditions are that the Euclidean distance is less than a preset distance and the time difference is less than a preset duration.

[0023] An electronic device, characterized in that it includes a processor and a storage device, the storage device stores multiple instructions, and the processor is used to read the instructions and execute the above method.

[0024] The multi-source data fusion method and device for road carbon emission monitoring provided by the present invention at least have the following beneficial effects: (1) Improve the richness of greenhouse gas monitoring information fusion. By fusing greenhouse gas data, moving vehicle position data, meteorological data, panoramic environmental image data, and environmental point cloud data, high-precision spatio-temporal matching is achieved, effectively eliminating the differences in spatial distribution and time points of a single data source, and ensuring the accuracy of urban road traffic greenhouse gas monitoring; (2) The spatio-temporal matching algorithm and multi-source data processing framework can effectively process data from different sensors and sources, provide reliable data support for further carbon emission estimation, and promote the implementation of intelligent transportation systems and green and low-carbon transportation policies; (3) Based on the nearest neighbor interpolation method and the inverse distance weighted method, data fusion in time and space is performed to improve the accuracy of data fusion. Description of the Drawings

[0025] Figure 1 It is a flowchart of an embodiment of a multi-source data fusion method for road carbon emission monitoring provided by the present invention.

[0026] Figure 2 It is a flowchart of an embodiment of data preprocessing in a multi-source data fusion method for road carbon emission monitoring provided by the present invention.

[0027] Figure 3 It is a flowchart of an embodiment of data matching in a multi-source data fusion method for road carbon emission monitoring provided by the present invention.

[0028] Figure 4 It is a flowchart of an embodiment of data matching using the nearest neighbor interpolation method in a multi-source data fusion method for road carbon emission monitoring provided by the present invention.

[0029] Figure 5 It is a flowchart of an embodiment of data matching using the inverse distance weighted method in a multi-source data fusion method for road carbon emission monitoring provided by the present invention.

[0030] Figure 6 It is a flowchart of an embodiment of a multi-source data fusion device for road carbon emission monitoring provided by the present invention. Detailed Embodiments

[0031] To better understand the above technical solutions, the following will describe the above technical solutions in detail in conjunction with the specification drawings and specific implementation manners.

[0032] Referring Figure 1 , in some embodiments, a multi-source data fusion method for road carbon emission monitoring is provided, including: S1. Collect multi-source carbon emission data based on the monitoring equipment on the mobile monitoring vehicle; S2. Preprocess the multi-source carbon emission data; S3. Construct an environmental model based on the preprocessed multi-source carbon emission data, and obtain the collection time of the multi-source carbon emission data. Obtain spatial information according to the constructed environmental model, and perform time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information; S4. Obtain the carbon emission spatio-temporal distribution result based on the multi-source carbon emission data that has been time-matched and space-matched.

[0033] Specifically, in step S1, a mobile monitoring system for urban roads is constructed by installing hardware systems such as greenhouse gas monitoring equipment, GPS system, panoramic camera, meteorological sensor, lidar measurement equipment, and data processor on the mobile monitoring vehicle. Based on the mobile monitoring system, greenhouse gas data, panoramic environmental image data, mobile vehicle position data, meteorological data, and environmental point cloud data can be obtained. Its main hardware devices include: Mobile monitoring vehicle: Select large vehicles such as SUVs as dedicated mobile vehicles, and perform brief modifications on them, including: setting the gas sampling pipeline inlet and fixing device; setting the power system fixing device, etc.; the vehicle body should be marked with significant signs such as mobile monitoring tests.

[0034] GPS system: The Global Positioning System (GPS) is essential for obtaining the mobile vehicle position data. At the same time, this system will provide accurate time, geographical coordinates, and vehicle driving status information, which is of great significance for the later analysis of mobile monitoring data.

[0035] Greenhouse gas measurement equipment: Used to detect and quantify the concentration of greenhouse gases in the atmosphere to obtain greenhouse gas data. According to requirements, in some embodiments, high-precision carbon dioxide (CO2), methane (CH4), etc. gas measurement equipment such as Picarro G2401, or medium-precision and low-cost CO2 monitoring equipment (Vaisala GMP343) can be used. Among them, the CO2 concentration measurement accuracy of the Picarro G2401 equipment is 0.1 ppm, and the measurement accuracy of the Vaisala GMP343 equipment is about 1-2 ppm.

[0036] Meteorological sensors: In addition to greenhouse gas data, meteorological data such as temperature, humidity, wind speed, and wind direction need to be collected. These data help to understand the distribution and changes of greenhouse gases in the atmosphere and carry out subsequent analysis and simulation of greenhouse gas concentrations.

[0037] Panoramic camera: To accurately reflect the driving of road vehicles and traffic flow conditions during data collection, a panoramic camera is installed in this system. The camera uses 5 fisheye cameras respectively for the front, rear, left, right, and up directions of the vehicle head, and has an image overlap of 40%-60%. Later, panoramic reconstruction can be carried out based on these 5 photos to obtain panoramic environmental image data.

[0038] LiDAR measurement equipment: The AS-900 equipment is installed, the scanning beam is adjusted to 200 revolutions per second, and the scanner frequency is set to 550KHZ to obtain environmental point cloud data of the road surface and roadside appurtenances, street trees, etc. on both sides of the road.

[0039] Data recording equipment: A data recording equipment is used to collect and store the data generated by the sensors. In some embodiments, a single-board Linux system processor such as a Raspberry Pi can be used, combined with a data SD memory card, to collect greenhouse gas data, moving vehicle position data, meteorological data, panoramic environmental image data, and environmental point cloud data per second, which is convenient for subsequent data processing.

[0040] Communication equipment: In order to enable real-time monitoring and remote control, a wireless network connection is required. In some embodiments, a 4G Internet of Things card of Huawei Cloud is used to transmit the data in the BeagleBone single-board Linux system processor to the laboratory server every 1 hour.

[0041] Power supply system: Since the monitoring equipment needs to run for a long time, and in addition, the equipment also requires high-power consumption situations such as constant-temperature heating, air pump decompression, and network data transmission. In some embodiments, a large mobile power supply (such as a mobile power supply of brands like Jackery, with a stored power of about 2 degrees of electricity, which can support the continuous power supply of the monitoring equipment for about 8 hours) can be selected to provide long-term and stable power supply for the observation equipment and other auxiliary equipment.

[0042] Safety equipment: To ensure the safety of vehicles and equipment, safety cameras can be set up.

[0043] Based on this, the above-mentioned multi-source carbon emission data includes greenhouse gas data, moving vehicle position data, meteorological data, panoramic environmental image data, and environmental point cloud data.

[0044] Further, referring to Figure 2 , in step S2, the preprocessing of the multi-source carbon emission data includes: S21. Calculate the standard deviation and mean for the same type of data in the multi-source carbon emission data; S22. Calculate the Z-score for each data point based on the standard deviation and mean, determine whether the data point is an outlier according to the Z-score, and delete the outlier; S23. Perform interpolation filling on the time-series data in the multi-source carbon emission data based on the linear interpolation method.

[0045] Specifically, in steps S21 and S22, during the data collection process, outliers may occur due to reasons such as sensor failures, measurement errors, or environmental interference. The existence of outliers will have an adverse impact on subsequent analysis, so it is necessary to screen them out. In this embodiment, the Z-score method is used to screen outliers. This method detects data points with large deviations from the mean by calculating the standard deviation of the data points. For the same type of data, that is, greenhouse gas data or meteorological data, a data point , its Z-score Z(xi) is calculated using the following formula: ; (1) where, is the mean of all data in the same type of data, is the standard deviation of all data in the same type of data.

[0046] In some embodiments, data points with a Z-score greater than the first preset value or less than the second preset value are considered outliers and are deleted. For example, the first preset value can be 3 and the second preset value can be -3.

[0047] Furthermore, in step S23, data missing is a common problem in the actual data collection process, which may be caused by sensor failures, communication interruptions, or external factors. Before performing data spatio-temporal matching, it is necessary to fill in the missing data. Methods such as linear interpolation can be used for data filling. Specifically, performing interpolation filling on the time-series data in the multi-source carbon emission data based on the linear interpolation method includes: Sort each type of time-series data according to the collection time respectively; Determine the missing points in the time-series data according to the collection frequency of each type of data; Calculate the value of the missing point based on the value of the previous data point and the value of the subsequent data point of the missing point.

[0048] Among them, the time-series data includes greenhouse gas data and meteorological data.

[0049] For time-series data, assuming that the change between data points is linear, the following formula is used for interpolation filling: ; (2) Among them, x(t) is the interpolation result of the missing point, x(t - 1) and x(t + n) are the values of the previous data point and the subsequent data point of the missing point respectively. Among them, n - 1 is the number of missing values.

[0050] Furthermore, referring to Figure 3 , in step S3, an environmental model is constructed based on the pre - processed multi - source carbon emission data, and the acquisition time of the multi - source carbon emission data is obtained. Spatial information is obtained based on the constructed environmental model, and time matching and spatial matching are performed on the multi - source carbon emission data based on the acquisition time and the spatial information, including: S31. Use a neural network model to identify vehicles in the panoramic environmental image data to obtain vehicle types and quantities; S32. According to the environmental point cloud data, construct an environmental three - dimensional point cloud model, and according to the environmental three - dimensional point cloud model, extract road vector data and form a first data set. The road vector data includes the road point cloud acquisition time and road position coordinates; S33. Resample the greenhouse gas data and meteorological data according to a preset time window to obtain greenhouse gas resampled data and meteorological resampled data under the time window; S34. According to the acquisition time of the panoramic environmental image data, greenhouse gas resampled data, and meteorological resampled data and the corresponding moving vehicle position data at that acquisition time, set timestamps and position coordinates for vehicle types and quantities, greenhouse gas resampled data, and meteorological resampled data and form a second data set; S35. Match the first data set and the second data set based on the nearest neighbor interpolation method or the inverse distance weighted method.

[0051] Specifically, in step S31, using a neural network model to identify vehicles in the panoramic environmental image data includes: Input the environmental image data into the pre - trained YOLOv8 for target recognition to obtain the types and quantities of target vehicles Specifically, input the panoramic environmental image data into the YOLOv8 model, scale it to a preset size at a certain ratio, usually 640x640 or 1280x1280, to ensure that the format of the input data is unified to adapt to the structure of the model. Then extract the features in the input image, from low - level edge and texture information to high - level object structure information, layer by layer to obtain the features in the image. In addition, YOLOv8 adds a Feature Pyramid Network (FPN) and a Path Aggregation Network (PAN) in this process to ensure that features at different scales can be effectively extracted to adapt to vehicles of different sizes.

[0052] Further, anchor box generation and prediction: YOLOv8 divides the entire image into grids, and each grid predicts multiple anchor boxes. Each anchor box contains the location information and class probabilities of possible objects. Estimate the position and size of the vehicle through the anchor box, and label each box with class labels such as "car", "truck", "bus", etc.

[0053] Further, use non-maximum suppression (NMS) technology to eliminate redundant detection results, retain the bounding box with the highest confidence, and suppress other boxes with an overlap ratio exceeding the threshold to ensure that each vehicle corresponds to only one detection box.

[0054] Further, classification and localization result output: Output a set of recognition results, and each result includes the class of the vehicle, the confidence score, and the bounding box position. Thus, the vehicle type and quantity are obtained.

[0055] Further, in step S32, a lidar measurement device is used to collect environmental point cloud data, and the spatial geometric texture and attribute information of the road are obtained, including road name, number of motor vehicle lanes, one-way or two-way traffic conditions, road type, etc.

[0056] Further, the environmental point cloud data is solved, the geometric features of the road are extracted, the point cloud data is matched to the correct geographic coordinate system using the measured correction points, and the least squares method is used to adjust the spatial relationship between the point cloud data and the correction points to reduce errors and ensure the accuracy and reliability of the data.

[0057] Based on the corrected environmental point cloud data, the vector elements of the road are extracted, mainly including the spatial position coordinates, shapes, and sizes of elements such as motor vehicle lanes, non-motor vehicle lanes, sidewalks, and green belts.

[0058] The collected three-dimensional vector elements are classified, integrated, and topologically checked to construct a complete environmental three-dimensional point cloud model, including motor vehicle lanes, non-motor vehicle lanes, sidewalks, green belts, etc. The road vector data is extracted and a first data set is formed. The road vector data includes the road point cloud collection time and the road position coordinates.

[0059] Further, in step S33, according to the preset time window, resample the greenhouse gas data and meteorological data, including: Calculate the mean value of the greenhouse gas concentration within the time window to obtain the greenhouse gas resampled data; Calculate the mean values of the temperature and humidity within the time window to obtain the meteorological resampled data.

[0060] Due to significant differences in the sampling times and time intervals of different sensors, it is necessary to perform time resampling on the sampling data of each sensor to ensure that all data has the same time granularity. For example, the sampling frequency of CO2 concentration data is once per second, while the onboard meteorological sensor has a time resolution of every 2 seconds. At the same time, the sampling frequency of traffic data and meteorological data in the later stage may be 1 minute. Therefore, in this embodiment, by setting a fixed time window (such as 60 seconds), the original data is resampled in time, and the synthetic data for the current time period is obtained through the arithmetic mean method. The specific formula (taking the monitoring data of atmospheric CO2 concentration as an example) is as follows: ; (3) Wherein, represents the CO2 concentration value after resampling processing at time t, , and represent the original time-resolution (1 second) CO2 concentration monitoring values within the time window at time t, and n is the number of monitoring data points within the time window.

[0061] Further, in step S34, according to the acquisition times of the panoramic environment image data, greenhouse gas resampling data, and meteorological resampling data, and the corresponding driving vehicle position data at the acquisition times, time stamps and position coordinates are set for the vehicle type and quantity, greenhouse gas resampling data, and meteorological resampling data, and a second data set is formed.

[0062] Further, referring to Figure 4 , in step S35, based on the nearest neighbor interpolation method, the first data set and the second data set are matched, including: S35a. Select the data points to be matched from the first data set as target points, and calculate the Euclidean distance and time difference between the target points and each data point in the second data set; S35b. Select the optimal data point from the data points in the second data set where the Euclidean distance and time difference meet the preset conditions, and associate it with the target point to obtain the matching result.

[0063] Specifically, a first data set Droad and a second data set Denv are constructed. Among them, the first data set Droad contains the road point cloud acquisition time and road position coordinates, and the second data set Denv is the data that needs to be matched with the road vector data, that is, the time stamps and the position coordinates at the time of acquisition corresponding to the vehicle type and quantity, greenhouse gas resampling data, and meteorological resampling data.

[0064] For each element i in the first dataset Droad and the second dataset Denv, define its position coordinates as two-dimensional coordinates (xi, yi), where xi and yi are the longitude and latitude respectively.

[0065] Convert each position coordinate into a geometric point object Pi = (xi, yi) for subsequent calculation use.

[0066] Define a preset duration △T to limit the maximum time difference that can be considered a match between two data points. For example, data within a duration of ±1 minute is considered a valid match.

[0067] Set a preset distance D, which represents the maximum distance (in meters) when performing a spatial join. Spatially, when the distance between two geometric objects is less than or equal to this threshold, they are considered a match. Temporally, the temporal proximity condition also needs to be satisfied.

[0068] Furthermore, select the data point Pa to be matched from the first dataset Droad as the target point, and calculate the Euclidean distance between it and any data point Pb in the second dataset using the following formula: ; (4) Assume that the acquisition time of the target point Pa is ta and the acquisition time of the data point Pb is tb. When d(Pa, Pb) ≤ D and ∣ta - tb∣ ≤ ΔT, Pa and Pb are considered potential matching pairs, that is, the preset condition is that the Euclidean distance is less than the preset distance and the time difference is less than the preset duration.

[0069] Specifically, for the target point Pa in the first dataset Droad, find all data points in the second dataset Denv that satisfy the conditions d(Pa, Pb) ≤ D and ∣ta - tb∣ ≤ ΔT, and select the data point with the minimum distance as the optimal data point. Match the attribute values of this optimal data point with the corresponding target point to obtain the matching result. If no qualified point is found, record it as empty or mark it as an unmatched state. Generate a new dataset based on the above matching result, which includes all columns of the original road vector data and the relevant information of the nearest neighbor found from the second dataset Denv. Remove all unmatched road records (i.e., those roads that fail to find a qualified environmental monitoring station) to obtain the final valid matching result.

[0070] Furthermore, refer to Figure 5 , in step S35, based on the inverse distance weighting method, match the first dataset and the second dataset, including: S351. Select the data points to be matched from the first dataset as target points, and calculate the Euclidean distance and time difference between each target point and each data point in the second dataset; S352. Set an initial search radius. For the target point, count the number of data points in the second dataset within the initial search radius whose Euclidean distance and time difference meet the preset conditions, and calculate the local point density; S353. Adjust the search radius according to the local point density and the preset density; S354. Select the data points in the second dataset that meet the preset conditions within the adjusted search radius, and perform distance-weighted calculation based on the attribute values of the selected data points and their Euclidean distance from the target point to obtain an attribute estimate value; S355. Associate the attribute estimate value with the target point to obtain a matching result.

[0071] Specifically, in step S351, select the data point Pa to be matched from the first dataset Droad as the target point, calculate its Euclidean distance from any data point Pb in the second dataset through formula (4), and calculate its time difference |ta - tb|.

[0072] Further, in step S352, set the initial search radius r0. For the target point Pa, according to the Euclidean distance, count the number n of data points in the first dataset within the initial search radius r0 whose Euclidean distance and time difference meet the preset conditions local , that is, the data points that satisfy d(Pa, Pb) ≤ D and |ta - tb| ≤ ΔT, and calculate the local point density density pa .

[0073] Among them, the initial search radius r0 can be one to two times the average distance between data points, or set according to the empirical value of the specific application scenario.

[0074] The local point density density pa is calculated by the following formula: ; (5) Among them, r0 represents the initial search radius, and n local represents the number of data points in the first dataset within the initial search radius r0 whose Euclidean distance and time difference meet the preset conditions.

[0075] Furthermore, in step S353, set a preset density d target , and this preset density can be an empirical value or the optimal density value determined through experiments. It represents the ideal local data point density, that is, the best interpolation effect can be obtained at this density.

[0076] According to the local point density density pa and the preset density d target Adjust the search radius: ; (6) where r a is the adjusted search radius.

[0077] If the local point density density Pa is greater than the preset density d target , then the adjusted search radius r a will be less than the initial search radius r0; conversely, if the local point density density Pa is less than the preset density d target , then the adjusted search radius r a will be greater than the initial search radius r0.

[0078] Furthermore, in step S354, assume that the number of data points belonging to the first data set that meet the preset conditions within the adjusted search radius is n, and perform distance-weighted calculation according to the following formula to obtain the attribute estimation value: ; (7) where Z(Pa) represents the attribute estimation value matching the target point Pa, zi represents the attribute value of the i-th data point Pbi belonging to the first data set that meets the preset conditions within the adjusted search radius, which can be the greenhouse gas concentration value, the number of vehicle types value, the meteorological data value, n represents the number of data points belonging to the first data set that meet the preset conditions within the adjusted search radius, p is the weight exponent, usually taking values between 1 and 3, used to control the influence degree of distance on the weight, and d(Pa, Pbi) represents the Euclidean distance between the target point Pa and the data point Pbi.

[0079] Specifically, in the above formula, the numerator: , which is the sum of the ratios of the attribute values of all eligible data points Pbi to the p-th power of their distances to the target point Pa. This means that the attribute value of each data point Pbi is weighted according to its distance to the target point, and the closer the point is to the target point, the greater the influence on the final estimation value.

[0080] The denominator: is to normalize each weighted value in the numerator to ensure that the finally obtained estimation value is within a reasonable range. It is the sum of the reciprocals of the p-th powers of the distances from all eligible data points Pbi to the target point Pa.

[0081] In step S355, associate the attribute estimate value with the target point to obtain a matching result.

[0082] Further, in step S4, establish a mapping relationship between the selected target points and the attribute values of the matched data points through time matching and space matching to obtain the carbon emission spatio-temporal distribution result.

[0083] Reference Figure 6 , in some embodiments, there is also provided a multi-source data fusion device for road carbon emission monitoring, including: A collection module 201 for collecting multi-source carbon emission data based on the monitoring equipment on the vehicle for driving survey; A preprocessing module 202 for preprocessing the multi-source carbon emission data; A matching module 203 for constructing an environmental model according to the preprocessed multi-source carbon emission data, obtaining the collection time of the multi-source carbon emission data, obtaining spatial information according to the constructed environmental model, and performing time matching and space matching on the multi-source carbon emission data based on the collection time and the spatial information; A result processing module 204 for obtaining the carbon emission spatio-temporal distribution result based on the multi-source carbon emission data with time matching and space matching.

[0084] Further, the multi-source carbon emission data includes greenhouse gas data, driving vehicle position data, meteorological data, panoramic environmental image data, and environmental point cloud data.

[0085] Further, the preprocessing module 202 preprocesses the multi-source carbon emission data, including: Calculating the standard deviation and mean value for the same type of data in the multi-source carbon emission data; Calculating the Z-score of each data point according to the standard deviation and mean value, determining whether the data point is an abnormal data point according to the Z-score, and deleting the abnormal data points; Performing interpolation filling on the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method.

[0086] Further, the preprocessing module 202 performs interpolation filling on the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method, including: Sorting each type of data belonging to the time series according to the collection time respectively; Determining the missing points in the data belonging to the time series according to the collection frequency of each type of data; Calculating the value of the missing point according to the value of the previous data point and the value of the subsequent data point of the missing point.

[0087] Further, the matching module 203 constructs an environmental model based on the preprocessed multi-source carbon emission data, obtains the collection time of the multi-source carbon emission data, obtains spatial information according to the constructed environmental model, and performs time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information, including: Use a neural network model to identify vehicles in the panoramic environmental image data to obtain vehicle types and quantities; According to the environmental point cloud data, construct an environmental three-dimensional point cloud model, and according to the environmental three-dimensional point cloud model, extract road vector data and form a first data set, where the road vector data includes the road point cloud collection time and road position coordinates; Resample the greenhouse gas data and meteorological data according to a preset time window to obtain greenhouse gas resampled data and meteorological resampled data under the time window; According to the acquisition time of the panoramic environmental image data, greenhouse gas resampled data, and meteorological resampled data and the corresponding driving vehicle position data at the acquisition time, set timestamps and position coordinates for the vehicle types and quantities, greenhouse gas resampled data, and meteorological resampled data and form a second data set; Match the first data set and the second data set based on the nearest neighbor interpolation method or the inverse distance weighted method.

[0088] Further, the matching module 203 matches the first data set and the second data set based on the nearest neighbor interpolation method, including: Select the data points to be matched from the first data set as target points, and calculate the Euclidean distance and time difference between the target points and each data point in the second data set; Select the optimal data point from the data points in the second data set where the Euclidean distance and time difference meet the preset conditions and associate it with the target point to obtain a matching result.

[0089] Further, the matching module 203 matches the first data set and the second data set based on the inverse distance weighted method, including: Select the data points to be matched from the first data set as target points, and calculate the Euclidean distance and time difference between the target points and each data point in the second data set; Set an initial search radius, for the target point, count the number of data points in the second data set within the initial search radius where the Euclidean distance and time difference meet the preset conditions, and calculate the local point density; Adjust the search radius according to the local point density preset density; Select data points belonging to the second data set that meet the preset conditions within the adjusted search radius, and perform distance-weighted calculation based on the attribute values of the selected data points and their Euclidean distances from the target point to obtain an attribute estimation value; Associate the attribute estimation value with the target point to obtain a matching result.

[0090] Further, the preset conditions are that the Euclidean distance is less than a preset distance and the time difference is less than a preset duration.

[0091] In some embodiments, an electronic device is also provided, including a processor and a storage device. The storage device stores multiple instructions, and the processor is configured to read the instructions and execute the above method.

[0092] The multi-source data fusion method and device for road carbon emission monitoring provided by the above embodiments have at least the following beneficial effects: (1) Improve the richness of greenhouse gas monitoring information fusion. By fusing greenhouse gas data, moving vehicle position data, meteorological data, panoramic environment image data, and environmental point cloud data, high-precision spatio-temporal matching is achieved, effectively eliminating the differences in spatial distribution and time points of a single data source, and ensuring the accuracy of urban road traffic greenhouse gas monitoring; (2) The spatio-temporal matching algorithm and the multi-source data processing framework can effectively process data from different sensors and sources, provide reliable data support for further carbon emission estimation, and promote the implementation of intelligent transportation systems and green and low-carbon transportation policies; (3) Based on the nearest neighbor interpolation method and the inverse distance weighting method, data is fused in time and space to improve the accuracy of data fusion.

[0093] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention. Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A multi-source data fusion method for road carbon emission monitoring, characterized in that Including: Collecting multi-source carbon emission data based on monitoring equipment on a mobile monitoring vehicle; Preprocessing the multi-source carbon emission data; Constructing an environmental model based on the preprocessed multi-source carbon emission data, obtaining the collection time of the multi-source carbon emission data, obtaining spatial information based on the constructed environmental model, and performing time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information; Obtaining a carbon emission spatio-temporal distribution result based on the multi-source carbon emission data that has been time-matched and space-matched.

2. The method according to claim 1, wherein The multi-source carbon emission data includes greenhouse gas data, mobile vehicle position data, meteorological data, panoramic environmental image data, and environmental point cloud data.

3. The method according to claim 1, wherein Preprocessing the multi-source carbon emission data includes: Calculating the standard deviation and mean for the same type of data in the multi-source carbon emission data; Calculating the Z-score for each data point according to the standard deviation and mean, determining whether the data point is an abnormal data point based on the Z-score, and deleting the abnormal data points; Performing interpolation filling on the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method.

4. The method according to claim 3, wherein Performing interpolation filling on the data belonging to the time series in the multi-source carbon emission data based on the linear interpolation method, including: Sorting each type of data belonging to the time series separately according to the collection time; Determining the missing points in the data belonging to the time series according to the collection frequency of each type of data; Calculating the value of the missing point according to the value of the previous data point and the value of the subsequent data point of the missing point.

5. The method according to claim 2, wherein Constructing an environmental model based on the preprocessed multi-source carbon emission data, obtaining the collection time of the multi-source carbon emission data, obtaining spatial information based on the constructed environmental model, and performing time matching and spatial matching on the multi-source carbon emission data based on the collection time and the spatial information, including: Identifying vehicles in the panoramic environmental image data using a neural network model to obtain the vehicle type and quantity; Constructing an environmental three-dimensional point cloud model based on the environmental point cloud data, and extracting road vector data from the environmental three-dimensional point cloud model to form a first data set, where the road vector data includes the road point cloud collection time and road position coordinates; Resampling the greenhouse gas data and meteorological data according to a preset time window to obtain greenhouse gas resampled data and meteorological resampled data under the time window; Setting time stamps and position coordinates for the vehicle type and quantity, greenhouse gas resampled data, and meteorological resampled data according to the acquisition time of the panoramic environmental image data, greenhouse gas resampled data, and meteorological resampled data and the corresponding mobile vehicle position data at that acquisition time to form a second data set; Matching the first data set and the second data set based on the nearest neighbor interpolation method or the inverse distance weighted method.

6. The method according to claim 5, wherein Matching the first data set and the second data set based on the nearest neighbor interpolation method, including: Selecting the data points to be matched from the first data set as target points, and calculating the Euclidean distance and time difference between the target points and each data point in the second data set; Select the optimal data point from the data points in the second data set where the Euclidean distance and time difference meet the preset conditions, and associate it with the target point to obtain a matching result.

7. The method according to claim 5, wherein Based on the inverse distance weighting method, matching the first data set and the second data set includes: Select the data point to be matched from the first data set as the target point, and calculate the Euclidean distance and time difference between the target point and each data point in the second data set; Set an initial search radius. For the target point, count the number of data points in the second data set where the Euclidean distance and time difference within the initial search radius meet the preset conditions, and calculate the local point density; Adjust the search radius according to the local point density and the preset density; Select the data points in the second data set that meet the preset conditions within the adjusted search radius, and perform distance weighting calculation based on the attribute values of the selected data points and their Euclidean distances from the target point to obtain an attribute estimation value; Associate the attribute estimation value with the target point to obtain a matching result.

8. The method according to claim 6 or 7, characterized in that, The preset conditions are that the Euclidean distance is less than a preset distance and the time difference is less than a preset duration.

9. A multi-source data fusion device for road carbon emission monitoring, characterized in that, It includes: An acquisition module for acquiring multi-source carbon emission data based on the monitoring devices on the vehicle during cruise monitoring; A preprocessing module for preprocessing the multi-source carbon emission data; A matching module for constructing an environmental model based on the preprocessed multi-source carbon emission data, obtaining the acquisition time of the multi-source carbon emission data, obtaining spatial information based on the constructed environmental model, and performing time matching and spatial matching on the multi-source carbon emission data based on the acquisition time and the spatial information; A result processing module for obtaining the spatio-temporal distribution result of carbon emissions based on the multi-source carbon emission data with time matching and spatial matching.

10. An electronic device, characterized in that, It includes a processor and a storage device. The storage device stores multiple instructions, and the processor is configured to read the instructions and execute the method according to any one of claims 1-8.

Images