Multi-source data fusion-based method for accurately searching high-emission vehicles on actual road

Abstract

The invention provides a multi-source data fusion-based actual road high-emission vehicle accurate searching method, and belongs to the field of motor vehicle pollution monitoring and emission control. Comprising the steps that remote sensing monitoring points are arranged on different road types to collect passing information of running vehicles, an old vehicle database, a static motor vehicle database and a vehicle large-user account are integrated, and a regional vehicle information database is established; carrying out trajectory analysis on the vehicle based on the database and the remote sensing data, extracting the trajectory length and the passing area of the high-emission vehicle, and analyzing the spatial-temporal distribution characteristics of the high-emission vehicle; in combination with remote sensing monitoring emission information and tracks, pollution contribution verification of high-emission vehicles in specific road sections and time sections is realized; and formulating classified management and control measures according to an expulsion result, and evaluating the emission reduction benefit. According to the invention, the problem that the existing motor vehicle management system cannot accurately master the activity condition of the high-emission vehicle can be solved, and technical support is provided for precise treatment.

Description

Technical Field

[0001] The present invention belongs to the field of motor vehicle pollution monitoring and emission control, and more specifically relates to a method for accurately mapping high-emission vehicles on actual roads based on multi-source data fusion. Background Art

[0002] With the continued increase in the number of motor vehicles, motor vehicle emissions have become a significant source of urban air pollution. To strengthen motor vehicle pollution prevention and control, existing motor vehicle management systems generally rely on static registration data, recording and supervising vehicle ownership and registration information. However, in practice, there is a disconnect between vehicle ownership and usage rights. For example, some vehicles may be registered locally but actually operate elsewhere, or they may be affiliated or operated on behalf of others. This makes it difficult to fully and dynamically capture the activity patterns and true emissions of vehicles in a specific area. Furthermore, traditional management systems often rely on surveys and inquiries with major vehicle users to obtain vehicle usage information. However, due to the predominantly manual data collection method, information updates are delayed, and are subject to subjectivity and incompleteness. This makes it difficult to fully and accurately reflect the actual driving conditions and emission characteristics of older, high-emission vehicles in a region. These issues directly impact the precise identification and effective supervision of high-emission vehicles, complicating air quality improvement and motor vehicle pollution prevention and control. Therefore, there is an urgent need for a technical means based on dynamic road remote sensing monitoring and multi-source data matching and fusion. By real-time monitoring and information integration of vehicles operating on actual roads, the operating characteristics and spatiotemporal distribution of high-emission vehicles in the region can be accurately grasped, thereby providing scientific and effective support for motor vehicle pollution control and policy formulation. Summary of the Invention

[0003] This paper proposes a precise method for mapping high-emission vehicles based on the fusion of multiple data sources, aiming to address the inaccuracy and incompleteness of traditional statistical methods. By integrating motor vehicle remote sensing data, road monitoring data, and registration databases, a high-precision database of high-emission vehicles is established, providing a scientific basis for the formulation of precise emission reduction strategies.

[0004] In order to achieve the above object, the present invention is implemented by adopting the following technical solutions: the method comprises:

[0005] Step 1: Data collection. This involves collecting vehicle traffic information, including travel time, vehicle type, and license plate information, from multiple motor vehicle remote sensing monitoring points on different types of roads within the region. This includes collecting data from a database of used vehicles, a static motor vehicle database, and major vehicle users. This data is then integrated with registered vehicle information within the region to establish a regional vehicle information database.

[0006] Step 2: Based on the database and remote sensing monitoring data, perform vehicle trajectory analysis, combine travel time and spatial distribution, extract the trajectory length and travel area range of high-emission vehicles, and analyze the spatiotemporal distribution characteristics of high-emission vehicles;

[0007] Step 3: Combine the high-emission vehicle information obtained by remote sensing monitoring equipment with its trajectory to verify the pollution contribution of such vehicles in time and space;

[0008] Step 4: Based on the results of the survey of high-emission vehicles, propose control measures, including traffic restrictions on key roads, upgrading emission standards, and eliminating subsidies.

[0009] In one scheme, the layout of the remote sensing monitoring points is based on the different road types, and is distributed according to the different road types in the region, such as main roads, secondary roads, branch roads and expressways. A random selection and continuous monitoring method is adopted to set up multiple monitoring points on various types of roads, and the vehicle travel time, vehicle type and license plate information of each monitoring point are collected according to the actual road traffic characteristics.

[0010] In one scheme, the standardized method for setting up remote sensing monitoring points comprehensively considers the regional road network density and traffic flow characteristics, formulates optimized point layout principles that are suitable for monitoring point coverage, and determines the spacing and distribution of monitoring points on different types of roads.

[0011] In one solution, the multi-source data of the old vehicle database, static motor vehicle database and large vehicle user ledger are integrated, and the regional vehicle information database established contains at least the vehicle's license plate number, vehicle type, registration time, vehicle ownership unit, emission standards and production time.

[0012] In one scheme, vehicle information and time series images obtained from remote sensing monitoring are analyzed. By extracting the time series of high-frequency target vehicles, the continuity of vehicle trajectories and spatial distribution characteristics are analyzed, and the specific trajectory length and passage area range of high-emission vehicles are delineated, achieving in-depth exploration of their temporal and spatial distribution characteristics.

[0013] In one solution, step 3 combines remote sensing emission information of high-emission vehicles with their travel trajectories, including a comprehensive evaluation of the frequency of vehicles' appearance on key roads and in key time periods, their spatial driving range, and the scope of their emission impact.

[0014] In one plan, by statistically analyzing the spatiotemporal distribution characteristics and emission trajectories of specific vehicles, the number of high-emission vehicles and changes in emission levels before and after the implementation of different control measures are quantitatively calculated.

[0015] In one plan, based on the precise survey results of high-emission vehicles, proposals were put forward, including targeted optimization of road restriction areas, suggestions for adjusting emission control standards, and subsidies for the elimination of old vehicles.

[0016] Beneficial effects of the present invention:

[0017] The comprehensive utilization of multiple data sources improves the accuracy of high-emission vehicle identification, ensuring data timeliness while balancing cost-effectiveness and computing power savings. It combines road monitoring data with remote sensing image analysis to enable dynamic tracking of high-emission vehicles. Using artificial intelligence algorithms, it improves vehicle identification and matching efficiency. This provides a scientific basis for regional high-emission vehicle control and facilitates targeted emissions reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Flow chart of the method of the present invention;

[0019] Figure 2 This is the spatial distribution diagram of PM2.5 emissions in a certain area at 6 o'clock in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate exemplary embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as those understood by those skilled in the art to which the present invention pertains. The terms used in the present specification are for the purpose of describing specific embodiments only and are not intended to limit the present invention. To facilitate understanding of the present invention, a more comprehensive description of the present invention will be provided below with reference to the accompanying drawings. Typical embodiments of the present invention are shown in the drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0022] like Figure 1 A method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion is shown, comprising the following steps:

[0023] Step 1. Data collection: Relying on multiple motor vehicle remote sensing monitoring points, collect traffic information of vehicles actually traveling on roads in the area.

[0024] Based on the measured results, a standardized method for point placement was proposed to achieve optimal monitoring results. Based on this standardized method, multiple points on different road types within the region were randomly selected for continuous monitoring, capturing vehicle travel time, vehicle type, and license plate information for each road type.

[0025] Collect old car databases and static motor vehicle databases, and integrate the data of major car users to obtain information on vehicles registered in the area and establish a regional vehicle information database.

[0026] Relying on the remote sensing monitoring technology of motor vehicles, a number of monitoring points are arranged on different types of roads in the region, such as main roads, secondary roads, branch roads, highways, etc., to realize the collection of real-time traffic information of vehicles. The principle of point layout is to ensure the comprehensiveness and accuracy of data coverage, while reducing the interference of redundant and complicated data on analysis. In order to optimize the point setting, we propose a standardized point layout method. First, according to the regional road network density D and traffic flow characteristics T, the monitoring point coverage C is comprehensively considered to design the optimal distance L between points. For any road network, let C = f(L,D,T), by calculating C opt = maxf(L,D,T) to determine the optimal point spacing. Generally, the initial recommendation for grid layout is To ensure uniform coverage within a unit area, for specific high-traffic road sections, the distance between points can be reduced by increasing the density of the deployment, adjusting it to L′=k·L, where k∈[0,1] is the adjustment factor.

[0027] During the continuous monitoring process, each point device will capture the instantaneous information of the vehicle passing through, including the passing time t i , vehicle model information, license plate number p i And exhaust emission spectral characteristics. The collected data can be expressed in the form of matrix M:

[0028]

[0029] Each row represents the monitoring record of a vehicle. In order to ensure the representativeness of the monitoring data, the monitoring time T 监测 The choice of is particularly important. Assuming that the spatiotemporal distribution of vehicles on a certain road conforms to the Poisson distribution, the statistical model of traffic flow can be used:

[0030] Q(t)=λ·e -λt

[0031] Where Q(t) is the vehicle flow per unit time, and λ is the flow intensity parameter. By adjusting the monitoring time T 监测 , to ensure that the sampling volume N in different time periods meets the distribution balance, that is: P(N 样本 |N 总 ) = uniform distribution.

[0032] At the same time, the collection and integration of static motor vehicle databases is the basis for building a regional vehicle information database. This process will integrate three types of data sources: old vehicle registration database, static motor vehicle registration database, and vehicle major user unit ledger. The identification of old vehicles depends on the vehicle's factory time T 出厂 and emission standards 阶段 , can be filtered by the condition t 出厂 <t 阈值 or e 阶段 ≤e 低标 The static motor vehicle database records the basic information of all registered vehicles, including license plate p i 、Vehicle type c i 、Fuel type i , Emission stage i The major vehicle user ledger supplements the ownership relationship between units with a large number of vehicles (such as logistics companies, bus groups, etc.) and their vehicles.

[0033] By removing duplicate data and matching, a regional vehicle information database is finally constructed. Its data structure can be expressed as an association table R 车辆 :

[0034]

[0035] Each row of data represents the information of a vehicle. i Matching in a static database can achieve the fusion of two types of information. This matching process is usually implemented using a hash table or a fast comparison algorithm to ensure efficiency and accuracy.

[0036] Ultimately, through the above-mentioned remote sensing monitoring, database collection and fusion, a comprehensive and accurate data foundation was laid for the subsequent high-emission vehicle analysis.

[0037] Step 2. Data processing and matching: The collected regional vehicle traffic information is processed by a computer program and matched with the vehicle information database to supplement information such as fuel type, emission stage (or factory date), vehicle operation nature and registered owner.

[0038] The temporal and spatial resolution of traffic flow can be customized based on vehicle travel time and the density of monitoring point settings. This allows for statistical analysis of traffic flow trends within a unit timeframe, with up to minute-level temporal accuracy supported. The standardized point setting method provides a spatial accuracy of 1 km, but by increasing the number of points, a spatial accuracy of 500 m can be achieved.

[0039] On the one hand, by coupling the emission factors specified in national standards or the local vehicle emission factors established through emission tests, the corresponding spatiotemporal distribution of emissions is obtained through calculation; on the other hand, high-frequency vehicles or high-emission old vehicles passing through the screening area are monitored, and the major vehicle users in the area are identified from the perspective of actual driving or pollutant emissions.

[0040] Further classification and refinement can be carried out through supplementary information such as fuel type, emission stage (or production date), vehicle operation nature and registered owner, so as to understand the traffic flow trend of different fuel types, emission stages and vehicle types and grasp the travel characteristics of specific vehicles.

[0041] First, the vehicle data set M collected from remote sensing monitoring points 通行 Regional Vehicle Information Database R 车辆 Perform automatic matching to complete the static attributes of the vehicle. The core of the matching is based on the unique license plate number p i As the primary key for association operations. 通行 Each pass record m in i =(p i ,t i ,c i , vehicle model features i , emission spectrum i ), find p i In R 车辆 The corresponding record r in i =(p i ,f i ,e i ,t 出厂 , affiliated units i ). The vehicle information expansion matrix M obtained after matching 扩展 Contains all completion fields:

[0042]

[0043] This process is usually accelerated using hash indexes or distribution-based matching algorithms, and the time complexity can be optimized to O(n).

[0044] After the matching is completed, the temporal and spatial resolution of the traffic flow is calculated based on the vehicle traffic information and the standardized layout density of the monitoring points. The time resolution Δt can be specified by the user, ranging from 1 hour to minute level (such as Δt = 1min), so as to facilitate the statistical analysis of the traffic flow change trend within a unit time. Let n ij Indicates that in the time resolution interval [t i ,t i +Δt] and the number of vehicles monitored at point j, the spatiotemporal distribution of the total traffic flow Q(t,x,y) can be expressed as:

[0045]

[0046] in S( x,y ) is the coverage coordinate ( x,y ) monitoring point set.

[0047] Spatial resolution is determined by the density of the monitoring points. Assuming the spacing between monitoring points is L, the basic spatial resolution is Δx = Δy = L. By increasing the density of the monitoring points, the spacing is reduced to L′ = k·L (where k∈[0,1] is an adjustment factor), achieving higher resolution (e.g., up to 500m). After all data is interpolated and gridded, a continuous spatiotemporal distribution map of vehicle distribution and traffic flow is generated.

[0048] In terms of emission accounting, the emission of each vehicle type is combined with traffic data by coupling the national standard emission factor or regional localized emission factor model. For a certain vehicle i, its actual driving emission E i It can be expressed as:

[0049]

[0050] Where F is the correction factor, which is used to correct the driving conditions in different regions; L i is the mileage of vehicle i. Total regional emission intensity E 区域 (t,x,y) is:

[0051]

[0052] in S( t,x,y ) Is the space-time range (t,x,y) The collection of vehicles inside.

[0053] In the identification and classification analysis of high-emission vehicles, by screening high-frequency vehicles, high-emission factor vehicles and old vehicles in the traffic data, we can accurately identify the large vehicle users who contribute the most to pollution. The screening of high-frequency vehicles can be done by counting the vehicle traffic frequency f 通行,i , select f 通行,i >θ f (The frequency threshold is θ f ) vehicles. The identification of high emission vehicles is based on the emission E i >θ E (The emission threshold is θ E ), while the screening of old vehicles is based on the factory time screening condition t 出厂,i <t 阈值 and low emission stage e i ≤e 低标 .

[0054] Finally, the extended data is analyzed in detail by fuel type (such as diesel, gasoline, natural gas, etc.), emission stage (such as National IV, National V, National VI), vehicle type (such as passenger car, truck, bus, etc.), and the traffic flow change trend Q of different types of vehicles is counted. c These data provide precise support for revealing the travel characteristics, spatiotemporal distribution patterns, and contributions to regional pollution of specific types of vehicles, laying the foundation for the formulation of subsequent precise governance strategies.

[0055] Step 3. Intelligent Analysis and Modeling: Remote sensing satellites or drones are used to collect multi-temporal, multi-angle, high-resolution images, which are then processed using a sliding window. A high-precision vehicle recognition model is employed, leveraging artificial intelligence technology to intelligently analyze remote sensing imagery from a specific time period. Combined with manual and semi-automatic annotation, vehicle targets are annotated to identify the number of vehicles in an area during the same time period, verifying the accuracy of high-emission vehicle identification.

[0056] First, high-resolution image data of multiple phases and angles within the region is collected by remote sensing satellites or drones. Assume that the remote sensing image data is in the form of a two-dimensional matrix I(x, y, t), where (x, y) is the pixel coordinate of the image in the two-dimensional plane and t is the imaging time. To improve computational efficiency and processing accuracy, a sliding window mechanism is used to partition the remote sensing image. Given a window size w×h and a window movement step size s, the image is divided into N sub-regions. Each sub-region P k is a submatrix of the original image:

[0057] P k =I(x k :x k +w,y k :y k +h,t),x k =x0+k x s,y k =y0+k y s

[0058] Where (k x ,k y ) are the indexes of the window in the x and y directions respectively. By sliding the window slice, the image is fully covered and the interference of target overlap is reduced.

[0059] In the segmented sub-images, a high-precision vehicle recognition model is used for target detection. The vehicle recognition model is usually based on a convolutional neural network (such as Faster R-CNN, YOLO) or the latest Transformer architecture (such as DETR). The input is each sub-image \(P_k\) and the output is the bounding box set B of all detected vehicle targets.i , where each bounding box B i =(x i ,y i ,w i ,h i ) represents the position and size of the target vehicle. The target detection process can be expressed as:

[0060]

[0061] Where F is the target detection network, θ is the network weight parameter, m k is the sub-image P k The number of vehicles identified in the test. While detecting, the model will further predict the probability distribution p of each target belonging to different vehicle types (such as passenger car, truck, bus) i =(p i,1 ,p i,2 ,...,p i,C ), where C is the total number of types, argmax(p i ) represents the classification result.

[0062] In order to improve the accuracy of annotation and the reliability of verification results, the manual and semi-automatic annotation processes are combined for verification. Based on the preliminary detection results of the model, the false detection and missed detection targets are corrected through manual intervention. Let the manually labeled target set be B i ', and define the intersection over union (IoU) of the two sets to measure the detection accuracy:

[0063]

[0064] Where B and B′ are the automatically detected and manually annotated object bounding boxes, respectively. When the IoU exceeds a threshold τ (usually set to a value such as 0.5), the detection result is considered valid. Furthermore, to improve processing efficiency, semi-supervised learning methods with minimal human intervention can be used to optimize model performance, such as by expanding the training set through pseudo-label generation strategies.

[0065] After completing the vehicle target number recognition, the number of identified vehicles N 车辆 (t) is compared with the vehicle traffic information Q(t,x,y) obtained in step 2, and the consistency of the two in time and space distribution is calculated. Define the consistency measure C 一致 for:

[0066]

[0067] where N 车辆 (t,x,y) is the number of vehicles identified by remote sensing images, and Q(t,x,y) is the predicted number of vehicles by the traffic flow model. 一致Above the threshold (τ C ), and considered the two to be a good match.

[0068] At the same time, in order to optimize the recognition accuracy of high-emission vehicles, the vehicle location, type and driving trajectory detected in the image can be combined to further match high-pollution emission vehicles (using the vehicle information matched in step 2). For high-frequency targets in the image time series, analyze their trajectory continuity and spatial distribution characteristics And extract the trajectory length L i And the range of the traffic area R i :

[0069]

[0070] Combine high emission vehicle information with trajectory T i Combined, verify its pollution contribution in time and space.

[0071] Step 4. Application and Emission Reduction Strategy Development: Based on the survey results, precise control measures for high-emission vehicles are proposed, including restrictions on key roads, emission standard upgrades, and subsidy phase-outs. By analyzing the spatiotemporal distribution characteristics of specific vehicles, the impact of different measures on specific vehicles can be determined and the corresponding emission reduction benefits calculated. This provides data support for government departments to optimize emission control policies and enhance regional air quality management capabilities.

[0072] First, a comprehensive review of the distribution, types, and behavioral patterns of high-emission vehicles, obtained through the aforementioned remote sensing monitoring, database matching, and intelligent analysis, is required. Leveraging the previously established vehicle traffic and attribute database, spatiotemporal distribution model, and remote sensing identification results, key vehicles with high emissions and high traffic frequency can be accurately screened and classified. This will then identify their travel concentration areas, peak travel times, and operational characteristics within the regional road network.

[0073] Based on these refined analysis results, targeted control measures can be formulated. For example, for traffic corridors and highly polluted roads with a high concentration of high-emission vehicles, it is recommended to implement special traffic restrictions, allowing only vehicles that meet certain emission standards to pass, or to adopt dynamic flow control for high-emission vehicles during specific periods. At the same time, based on the vehicle's emission stage, usage nature, and cumulative traffic behavior, scientifically plan subsidy incentive policies for the early or time-limited elimination of old high-emission vehicles, clarify the list of vehicles that can be eliminated, and encourage car owners to accelerate vehicle replacement. For large vehicle-user groups and high-polluting corporate vehicles, differentiated supervision and dynamic monitoring can be implemented to encourage the use of new energy vehicles to replace traditional fuel vehicles, and gradually raise the environmental protection threshold for corporate vehicles.

[0074] Furthermore, policy simulations of various control measures are applied to target vehicles in the database, and their spatiotemporal distribution characteristics and traffic behavior are used to quantitatively evaluate their effectiveness. For example, after implementing a certain type of traffic restriction, it is possible to predict the number of high-emission vehicles in a region and the corresponding reduction in emissions during a specific time period, allowing for the scientific calculation of the environmental benefits of each measure. Such simulations and effect pre-evaluations not only facilitate dynamic fine-tuning of policy content and optimal solution selection, but also provide solid data support for relevant government departments in formulating and adjusting emission control policies and allocating environmental protection funds.

[0075] Ultimately, through this closed-loop mechanism, a big data-driven, precise and efficient high-emission vehicle management system will be gradually established, greatly improving the scientificity and precision of regional air quality management and promoting the goal of continuous improvement of ambient air quality.

[0076] Example:

[0077] For areas within a 5-kilometer radius, there should be no less than 3 monitoring points to reflect changes in traffic flow in the area; for areas larger than 5 kilometers, the traffic flow of regional roads should be monitored according to the standardized point setting method shown in Table 1 and the actual road grade in the area. This method takes into account the traffic flow conditions of different road types and can more accurately reflect changes in regional traffic flow.

[0078] Monitoring points should be randomly distributed across the road length. Traffic volume varies across different types of roads. Expressways and urban freeways typically have higher speeds and heavier traffic, requiring a higher density of monitoring points. Lower-level highways, on the other hand, have relatively lower traffic volumes, so monitoring points can be more sparsely distributed. Key entrances, exits, and toll booths also warrant special attention. Furthermore, economic and technical feasibility must be considered to ensure both coverage and rational deployment.

[0079] like Figure 2 As shown in the figure, taking a certain district as an example, 78 monitoring points were set up according to the standardized method, and the PM caused by road traffic in a certain area was finally measured at 6 o'clock. 2.5 The emission is 2.7 kg, the spatial resolution is 1 km × 1 km, the spatial distribution of the unit space area is shown in the figure, and the maximum emission intensity is 35 g / h.

[0080] Through intelligent analysis of satellite images, the number of motor vehicles on the road at 6 o'clock in a certain area deviated from the traffic flow monitoring results by 6%, and the overall consistency of the results was good.

[0081] Table 1 Standardized setting method for traffic flow monitoring points

[0082]

[0083] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0084] It should be understood that the detailed description of the technical solutions of the present invention using the preferred embodiments above is illustrative and not restrictive. A person skilled in the art, after reading the present specification, may modify the technical solutions described in the embodiments or replace some of the technical features therein with equivalents; such modifications or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion, characterized by: The method includes: Step 1: Data collection. This involves collecting vehicle traffic information, including travel time, vehicle type, and license plate information, from multiple motor vehicle remote sensing monitoring points on different types of roads within the region. This includes collecting data from a database of used vehicles, a static motor vehicle database, and major vehicle users. This data is then integrated with registered vehicle information within the region to establish a regional vehicle information database. Step 2: Based on the database and remote sensing monitoring data, perform vehicle trajectory analysis, combine travel time and spatial distribution, extract the trajectory length and travel area range of high-emission vehicles, and analyze the spatiotemporal distribution characteristics of high-emission vehicles; Step 3: Combine the high-emission vehicle information obtained by remote sensing monitoring equipment with its trajectory to verify the pollution contribution of such vehicles in time and space; Step 4: Based on the results of the survey of high-emission vehicles, propose control measures, including traffic restrictions on key roads, upgrading emission standards, and eliminating subsidies.

2. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1, characterized in that: The layout of the remote sensing monitoring points is based on the different road types, and is distributed according to the different road types of main roads, secondary roads, branch roads and expressways in the region. A random selection and continuous monitoring method is adopted to set up multiple monitoring points on various types of roads, and the vehicle travel time, vehicle type and license plate information of each monitoring point are collected according to the actual road traffic characteristics.

3. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1 is characterized by: The monitoring point setting of the remote sensing monitoring data is based on a comprehensive consideration of regional road network density and traffic flow characteristics, and an optimized point layout principle adapted to the monitoring point coverage is formulated to determine the spacing and distribution of monitoring points for different types of roads.

4. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1 is characterized by: The old car database, static motor vehicle database and large vehicle user ledger are integrated with multi-source data, and the established regional vehicle information database contains at least the vehicle's license plate number, vehicle type, registration time, vehicle ownership unit, emission standards and production time.

5. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1 is characterized by: By analyzing vehicle information and time series images obtained through remote sensing monitoring, and extracting the time series of high-frequency target vehicles, the continuity of vehicle trajectories and spatial distribution characteristics are analyzed, and the specific trajectory length and passage area range of high-emission vehicles are delineated, so as to achieve in-depth exploration of their spatiotemporal distribution characteristics.

6. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1 is characterized by: In step 3, the remote sensing emission information of high-emission vehicles is combined with their travel trajectories, including a comprehensive evaluation of the frequency of vehicles' appearance on key roads and in key time periods, their spatial driving range, and the scope of emission impact.

7. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1, characterized in that: By statistically analyzing the spatiotemporal distribution characteristics and emission trajectories of specific vehicles, the number of high-emission vehicles and changes in emission levels before and after the implementation of different control measures are quantitatively calculated.

8. The method for accurately identifying high-emission vehicles on actual roads based on multi-source data fusion according to claim 1 is characterized by: Based on the precise survey results of high-emission vehicles, proposals are put forward, including targeted optimization of road restriction areas, adjustment suggestions for emission control standards, and subsidy plans for the elimination of old vehicles.

Images