Large-cargo transport vehicle identification method and system based on radar and vision fusion

Abstract

The present invention relates to the field of data processing technology, and in particular to a method and system for identifying large-scale transport vehicles based on radar and vision fusion. The method comprises the following steps: collecting radar signals for monitoring the transport vehicle on the way through radar monitoring equipment; analyzing the transport vehicle characteristics of the radar signals based on the radar signals for monitoring the transport vehicle on the way, and generating transport vehicle characteristic data; collecting visual data on the way of the transport vehicle using high-resolution monitoring equipment; analyzing the visual features of the transport vehicle on the way based on the visual data on the way of the transport vehicle, and generating visual feature data on the way of the transport vehicle; mapping the external heterogeneous data of the transport vehicle and identifying the dynamic feature patterns of the transport vehicle based on the transport vehicle characteristic data and the visual feature data on the way of the transport vehicle, and generating the external heterogeneous data matrix of the transport vehicle and the dynamic feature pattern data of the transport vehicle, respectively. The present application identifies large-scale transport vehicles based on radar and vision fusion, and more comprehensively identifies the vehicle information of large-scale transport vehicles.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for identifying large-scale transport vehicles based on radar and vision fusion. Background Art

[0002] Identification methods for large-scale transport vehicles are critical to ensuring road safety and optimizing traffic flow. These vehicles usually have a significant impact on road structure and traffic flow due to the special size and weight. Effective identification of these vehicles helps traffic management systems monitor their driving status and location in real time, prevent potential traffic accidents, and develop more effective traffic planning and route optimization strategies. In addition, accurate vehicle identification also helps to enforce transportation regulations and ensure that heavy-loaded and oversized vehicles comply with specific driving routes and time limits, thereby reducing damage to road infrastructure and congestion problems. However, traditional vehicle identification methods for large-scale transport vehicles often identify transport vehicles through a single detection such as visual recognition. The large-scale transport vehicles are too large, and it is difficult to fully identify large-scale vehicles through a single camera, resulting in poor identification of transport vehicles. In addition, it is impossible to capture the dynamic pattern of the real-time location and status of large-scale transport vehicles, and it is impossible to provide more identification information of large-scale transport vehicles. Summary of the invention

[0003] Based on this, the present invention provides a method and system for identifying large-scale transport vehicles based on radar and vision fusion to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a large-scale transport vehicle identification method based on radar and vision fusion includes the following steps:

[0005] Step S1: Acquire the data of the supervision area on the way of the large-scale transport vehicle; collect the monitoring radar signal of the transport vehicle on the way through the radar monitoring equipment and the data of the supervision area on the way of the large-scale transport vehicle to generate the monitoring radar signal on the way of the transport vehicle; analyze the transport vehicle characteristics of the radar signal according to the monitoring radar signal on the way of the transport vehicle to generate the transport vehicle characteristic data, wherein the transport vehicle characteristic data includes the static geometry data of the transport vehicle and the kinematic data of the transport vehicle;

[0006] Step S2: using high-resolution monitoring equipment and data from the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle; performing visual feature analysis on the way of the transport vehicle based on the visual data on the way of the transport vehicle, and generating visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static appearance feature data of the transport vehicle and transport feature data of the transport vehicle;

[0007] Step S3: mapping the external heterogeneous data of the transport vehicle according to the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate a matrix of the external heterogeneous data of the transport vehicle;

[0008] Step S4: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle to generate the transport vehicle dynamic feature pattern data;

[0009] Step S5: Transmitting the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform vehicle identification feedback operation of the large-scale transport vehicle.

[0010] The radar technology of the present invention can work effectively in any weather conditions, including at night and in low visibility environments. This ensures that the monitoring of transport vehicles will not be interrupted regardless of environmental conditions, thereby improving the continuity and reliability of safety supervision. Radar signals can accurately measure the speed of transport vehicles and their relative positions to other objects, which is crucial to avoiding traffic congestion and accidents, especially on highways and in dense traffic environments. By analyzing radar signals, not only can the vehicle's motion state (such as speed and acceleration) data be obtained, but also the vehicle's static parameters (such as vehicle length and width) can be measured, providing a basis for further safety assessment and behavior analysis. High-resolution cameras provide clear visual images and can capture detailed features of vehicles, such as license plates, logos and vehicle colors, which are crucial to ensuring the accuracy and reliability of vehicle identification. Not only can the appearance features of the vehicle be extracted from the visual data, but the actual transportation behavior of the vehicle (such as loading status and loading and unloading process) can also be observed, which is conducive to more in-depth behavior analysis and risk assessment, and instant feedback on the status and behavior of the transport vehicle can be provided to support emergency response and immediate decision-making. Combining the static geometric data of the transporter (such as size and shape) with the appearance feature data (such as color and logo) to generate a comprehensive data matrix through heterogeneous data mapping, such fusion provides a comprehensive view of vehicle characteristics, which helps to more accurately identify and classify vehicles. Through detailed data level integration, richer information is provided to reduce misjudgment in the identification process, especially in the case of diverse vehicle types and models. The kinematic data (such as speed, acceleration) and transportation characteristics (such as loading status changes) of the transporter are analyzed to identify the behavior patterns of the vehicle. This is very useful for predicting vehicle behavior, assessing safety risks, and optimizing route planning. By monitoring the dynamic characteristics of the vehicle in real time, the system can instantly detect abnormal behavior, such as sudden acceleration or illegal lane changes, and take necessary safety measures. The timely transmission of integrated vehicle data to the terminal and the rapid execution of the identification and feedback process are crucial to the responsiveness and efficiency of the traffic management system. By using detailed vehicle characteristics and dynamic data, the management system can more effectively carry out traffic scheduling, route optimization and safety management, especially in complex situations involving large-scale transportation.

[0011] Preferably, step S1 comprises the following steps:

[0012] Step S11: Obtaining data of the supervision area on the way of the large-scale transport vehicle;

[0013] Step S12: collecting radar signals for monitoring the transport vehicle en route through radar monitoring equipment and the data of the supervision area for the large-scale transport vehicle en route to generate radar signals for monitoring the transport vehicle en route, wherein the radar signals for monitoring the large-scale transport vehicle en route include static size monitoring radar signals and dynamic kinematic monitoring radar signals;

[0014] Step S13: extracting the amplitude peak value according to the static size monitoring radar signal to generate an amplitude peak size monitoring radar signal;

[0015] Step S14: analyzing static geometric data of the transport vehicle according to the amplitude peak size monitoring radar signal to generate static geometric data of the transport vehicle;

[0016] Step S15: Performing kinematic data analysis on the transport vehicle based on the dynamic kinematic monitoring radar signal to obtain the kinematic data of the transport vehicle.

[0017] By clearly setting the supervision area, the present invention can ensure the pertinence and efficiency of data collection, and provide accurate geographical and environmental background information for subsequent steps. The collection of radar signals not only covers the static size information of the vehicle, but also includes dynamic kinematic information, thereby realizing all-round monitoring of the transport vehicle. By extracting the amplitude peak of the radar signal, the key geometric features of the vehicle, such as vehicle length, vehicle width, etc., can be accurately identified, which are important parameters for evaluating vehicle size and category. Analysis of the amplitude peak size monitoring radar signal can generate detailed static geometric data of the vehicle and provide accurate information on the vehicle structure, which is key to compliance inspection and classification of transport vehicles. Based on the analysis of dynamic kinematic monitoring radar signals, kinematic data such as vehicle speed and acceleration can be obtained, which is very useful for evaluating vehicle driving behavior, safety, and adjusting traffic management strategies.

[0018] Preferably, step S15 comprises the following steps:

[0019] Step S151: performing signal receiving angle analysis according to the dynamic kinematics monitoring radar signal to generate signal receiving angle data, and performing transport vehicle angle analysis through the signal receiving angle data to generate transport vehicle angle data;

[0020] Step S152: performing signal reception time difference measurement and analysis according to the dynamic kinematics monitoring radar signal to generate signal reception time difference data, and performing transport vehicle path data evaluation processing through the signal reception time difference data and the signal reception angle data to generate transport vehicle path data evaluation data;

[0021] Step S153: performing signal frequency change analysis according to the dynamic kinematics monitoring radar signal to generate signal frequency change analysis data, and performing Doppler effect transport vehicle speed analysis through the signal frequency change analysis data, signal reception time difference data, and signal reception angle data to generate transport vehicle speed data;

[0022] Step S154: Integrate the transport vehicle angle data, the transport vehicle path data evaluation data and the transport vehicle speed data to generate the transport vehicle kinematic data.

[0023] The present invention accurately measures and determines the direction and angle of the transport vehicle by analyzing the received signal angle, which is critical for understanding the vehicle's driving direction and its specific position on the road. By measuring the time difference of signal reception, the distance between the transport vehicle and the radar device is accurately evaluated. This provides accurate vehicle location information for traffic management, helps prevent collisions and optimize traffic flow. By analyzing the changes in signal frequency, the speed of the transport vehicle can be accurately measured using the Doppler effect. This is crucial for monitoring the vehicle's driving speed, predicting traffic flow changes, and enforcing speed limit policies. The transport vehicle's angle data, path data, and speed data are integrated to generate a comprehensive kinematic data set. This data integration provides a panoramic view of the transport vehicle's behavior, enhances the understanding of its dynamic performance, and supports more accurate behavior prediction and safety analysis.

[0024] Preferably, step S2 comprises the following steps:

[0025] Step S21: using high-resolution monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle, wherein the visual data on the way of the transport vehicle includes static transport vehicle image data and dynamic transport vehicle image data;

[0026] Step S22: extracting the appearance features of the static transport vehicle according to the static transport vehicle image data to generate the appearance feature data of the static transport vehicle;

[0027] Step S23: Analyze the transport characteristics of the transport vehicle based on the dynamic transport vehicle image data to generate transport characteristic data of the transport vehicle.

[0028] The present invention uses high-resolution monitoring equipment to obtain visual data of the transport vehicle, including static and dynamic images. This high-quality image acquisition provides detailed visual evidence for subsequent in-depth analysis, enhancing the availability of data and the depth of analysis. The data collected in the supervision area on the way of the large-scale transport vehicle not only records the behavior of the vehicle, but also captures the situation of the surrounding environment, which helps to more comprehensively evaluate the operating environment of the vehicle and possible interactions or conflicts. The appearance features of the transport vehicle, such as model, color, license plate, etc., are extracted from the static image. These detailed appearance feature data are crucial for the confirmation and classification of vehicle identity, especially in situations where different vehicles or vehicle types need to be accurately identified. By accurately extracting the static features of each vehicle, the error in the vehicle identification process is reduced, and the reliability and accuracy of the overall monitoring system are improved. Based on the dynamic image data of the transport vehicle, analyzing its transportation characteristics, such as the vehicle's driving speed, driving situation, load changes, etc., is extremely important for understanding the behavior pattern of the vehicle during actual transportation. Through in-depth analysis of the transportation characteristics of the transport vehicle, potential safety issues or efficiency bottlenecks can be identified, and measures can be taken to optimize the transportation route or improve the safety of transportation operations.

[0029] Preferably, step S23 includes the following steps:

[0030] Step S231: performing image binarization processing on the dynamic transport vehicle image data to generate binary dynamic transport vehicle image data;

[0031] Step S232: performing image segmentation according to the binary dynamic transport vehicle image data to generate segmented dynamic transport vehicle image data;

[0032] Step S233: analyzing the motion blur distortion parameters of the transport vehicle according to the segmented dynamic transport vehicle image data to generate motion blur distortion parameters;

[0033] Step S234: performing motion blur distortion correction processing on the dynamic transport vehicle image data according to the motion blur distortion parameters to generate corrected dynamic transport vehicle image data;

[0034] Step S235: Performing transportation characteristic analysis based on the corrected dynamic transport vehicle image data to generate transport characteristic data of the transport vehicle.

[0035] The present invention simplifies the complexity of the image by converting the dynamic transport vehicle image data into a binary format, making the subsequent processing steps more efficient and accurate, especially in image segmentation and feature extraction. The segmentation of the binary image makes it easier to accurately identify the outline and other important parts of the transport vehicle, which helps to effectively separate the vehicle from the complex background and improve the accuracy of target detection. Accurately analyzing the blur distortion parameters caused by vehicle motion provides necessary information for subsequent image correction, so that the image quality can be improved in a targeted manner. Applying motion blur distortion parameters for image correction effectively reduces or eliminates the blur caused by the rapid movement of the vehicle, restores the clarity of the image, and helps to perform subsequent feature analysis more accurately. The corrected image data provides clearer visual information, making the transportation feature analysis of the transport vehicle more accurate, such as vehicle load status, loading and unloading activities and behavior patterns, etc., which are crucial for transportation supervision and safety analysis.

[0036] Preferably, step S233 includes the following steps:

[0037] The middle nodes of the segmented dynamic transport vehicle image data are selected as the dynamic transport vehicle mass point parameters, and the dynamic transport vehicle mass point parameters are analyzed in time series to generate the time series dynamic transport vehicle mass point parameters;

[0038] Performing fuzzy area recognition on the segmented dynamic transport vehicle image data to generate dynamic transport vehicle fuzzy area data;

[0039] According to the time series dynamic transport vehicle particle parameters and the dynamic transport vehicle blur area data, the motion blur distortion parameters of the transport vehicle are analyzed to generate motion blur distortion parameters.

[0040] The present invention selects the middle nodes in the image data as particles and performs time series analysis on them, which can accurately track the movement path and speed of the transport vehicle, help better understand the specific movement state of the vehicle during monitoring, and provide key data for the motion model. The segmented dynamic transport vehicle image data is subjected to fuzzy area identification to clarify the key areas of image processing, that is, those fuzzy areas caused by rapid movement, which helps to make subsequent image correction more targeted and improve processing efficiency and effect. Combined with the time series dynamic transport vehicle particle parameters and the dynamic transport vehicle fuzzy area data, motion blur distortion parameter analysis is performed, which can accurately identify the specific blur type and degree caused by the rapid movement of the vehicle, so that the subsequent image correction work can be carried out based on the actual distortion situation, greatly improving the image availability and analysis accuracy.

[0041] Preferably, step S3 comprises the following steps:

[0042] Step S31: extracting the external information type of the transport vehicle from the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate the external information type of the transport vehicle;

[0043] Step S32: performing heterogeneous matrix architecture analysis according to the external information type of the transport vehicle to construct a heterogeneous matrix architecture;

[0044] Step S33: transmitting the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to the heterogeneous matrix architecture for data mapping of the external heterogeneous data of the transport vehicle to generate an external heterogeneous data matrix of the transport vehicle.

[0045] The present invention can more effectively classify and standardize this information by extracting key information types from the static geometric data and static appearance feature data of the transport vehicle, which not only helps to organize and manage the data, but also improves the efficiency and accuracy of subsequent data processing steps. Heterogeneous matrix architecture analysis is performed based on the extracted external information types of the transport vehicle, aiming to build a data architecture that can effectively process and store this information, optimize data query, storage and access efficiency, and provide a basis for more complex data analysis and decision support. The static geometric data and appearance feature data of the transport vehicle are mapped to the constructed heterogeneous matrix architecture to generate a transport vehicle external heterogeneous data matrix, which can provide a full view of vehicle characteristics and greatly enhance the application value of the data, especially when performing complex vehicle identification and classification tasks.

[0046] Preferably, step S32 includes the following steps:

[0047] Perform unique identification external information type analysis based on the external information type of the transport vehicle to generate unique identification external information type data;

[0048] Designing unique identification nodes of heterogeneous matrices by uniquely identifying external information type data, and designing heterogeneous matrix nodes by transport vehicle external information type;

[0049] Perform heterogeneous matrix node association processing on the heterogeneous matrix unique identification node and the heterogeneous matrix node to generate heterogeneous matrix node association data;

[0050] A heterogeneous matrix architecture is established for the unique identification nodes of the heterogeneous matrix and the heterogeneous matrix nodes according to the heterogeneous matrix node association data to construct a heterogeneous matrix architecture.

[0051] The present invention ensures that each type of data has a clear identifier by performing unique identification analysis on the external information type of the transport vehicle, which helps to avoid data duplication and errors and increase the overall accuracy and reliability of the database. The unique identification node in the heterogeneous matrix is ​​designed using the unique identification external information type data, which improves the logic and search efficiency of the data structure and makes data retrieval faster and more accurate. By associating the unique identification node and other nodes of the heterogeneous matrix, a complex data relationship network is created, which not only helps to understand the relationship between the data, but also improves the depth and breadth of data analysis. The heterogeneous matrix architecture is constructed based on the node association data, providing a highly flexible and scalable platform that supports various types of data analysis and processing, not only meeting current needs, but also easily adapting to future expansion and modification.

[0052] Preferably, step S4 comprises the following steps:

[0053] Step S41: integrating the transport vehicle kinematic data and the transport vehicle transport characteristic data into dynamic heterogeneous transport vehicle data;

[0054] Step S42: performing dynamic key frame analysis based on the dynamic heterogeneous data of the transport vehicle to generate key frame dynamic heterogeneous data of the transport vehicle;

[0055] Step S43: performing time series analysis on the key frame transport vehicle dynamic heterogeneous data to generate time series key frame transport vehicle dynamic heterogeneous data, and performing nonlinear transport vehicle dynamic spatiotemporal feature analysis based on the time series key frame transport vehicle dynamic heterogeneous data to generate transport vehicle dynamic spatiotemporal feature data;

[0056] Step S44: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle dynamic spatiotemporal feature data to generate transport vehicle dynamic feature pattern data.

[0057] The present invention integrates the kinematic data and transportation characteristic data of the transport vehicle into a unified data set, providing a comprehensive perspective on the dynamic behavior of the transport vehicle, and increasing the consistency of data processing and the depth of analysis. By analyzing and determining the key frames in the dynamic data, important moments and turning points in the transportation process are identified, which is crucial to understanding the behavior of the transport vehicle and making corresponding responses. Time series analysis is performed on the key frame dynamic heterogeneous data, and further nonlinear dynamic spatiotemporal characteristic analysis is performed to reveal the behavior patterns and dynamic changes of the transport vehicle, which are very valuable for predicting future behavior and conducting risk assessments. Based on the dynamic spatiotemporal characteristic data of the transport vehicle, dynamic characteristic patterns are identified to effectively predict the possible behavioral trends and risks of the transport vehicle, support decision-making and emergency response plans, and thus improve transportation safety and efficiency.

[0058] This specification also provides a large-scale transport vehicle identification system based on radar and vision fusion, which is used to execute the large-scale transport vehicle identification method based on radar and vision fusion as described above. The large-scale transport vehicle identification system based on radar and vision fusion includes:

[0059] The radar signal transport vehicle characteristic analysis module is used to obtain the data of the supervision area on the way of the large-scale transport vehicle; collect the transport vehicle monitoring radar signal on the way through the radar monitoring equipment and the data of the supervision area on the way of the large-scale transport vehicle to generate the transport vehicle monitoring radar signal on the way; analyze the transport vehicle characteristics of the radar signal according to the transport vehicle monitoring radar signal on the way to generate the transport vehicle characteristic data, wherein the transport vehicle characteristic data includes the transport vehicle static geometry data and the transport vehicle kinematic data;

[0060] The transport vehicle visual feature analysis module is used to collect visual data of the transport vehicle on the way by using high-resolution monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle, and generate visual data of the transport vehicle on the way; perform visual feature analysis on the transport vehicle on the way according to the visual data on the way of the transport vehicle, and generate visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static transport vehicle appearance feature data and transport feature data of the transport vehicle;

[0061] A transport vehicle external heterogeneous data mapping module is used to map the transport vehicle external heterogeneous data according to the transport vehicle static geometric data and static transport vehicle appearance feature data to generate a transport vehicle external heterogeneous data matrix;

[0062] The transport vehicle dynamic feature pattern recognition module is used to perform transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle, and generate the transport vehicle dynamic feature pattern data;

[0063] The transport vehicle identification feedback module is used to transmit the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform the vehicle identification feedback operation of the large-scale transport vehicle.

[0064] The beneficial effect of the present application is that the method for identifying large-scale transport vehicles based on radar and vision fusion of the present invention utilizes the ability of radar to provide accurate distance and speed data under any weather and lighting conditions, and combines the detailed image and appearance information provided by the visual system to ensure high-precision vehicle identification and classification under various environmental conditions. Through in-depth analysis of kinematic data and transportation characteristics, the identification method can not only capture the vehicle's real-time position and status, but also predict the risk of large-scale transport vehicles through timing and dynamic feature analysis, providing support for traffic safety and route planning. And through the fusion analysis of radar and vision, more identification information of large-scale transport vehicles is provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A schematic diagram of the steps of a method for identifying large-scale transport vehicles based on radar and vision fusion according to the present invention;

[0066] Figure 2 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0067] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0068] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0069] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0070] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly a second unit may be referred to as a first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0071] To achieve this, please refer to Figure 1 to Figure 2 The present invention provides a method for identifying large-scale transport vehicles based on radar and vision fusion, comprising the following steps:

[0072] Step S1: Acquire the data of the supervision area on the way of the large-scale transport vehicle; collect the monitoring radar signal of the transport vehicle on the way through the radar monitoring equipment and the data of the supervision area on the way of the large-scale transport vehicle to generate the monitoring radar signal on the way of the transport vehicle; analyze the transport vehicle characteristics of the radar signal according to the monitoring radar signal on the way of the transport vehicle to generate the transport vehicle characteristic data, wherein the transport vehicle characteristic data includes the static geometry data of the transport vehicle and the kinematic data of the transport vehicle;

[0073] Step S2: using high-resolution monitoring equipment and data from the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle; performing visual feature analysis on the way of the transport vehicle based on the visual data on the way of the transport vehicle, and generating visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static appearance feature data of the transport vehicle and transport feature data of the transport vehicle;

[0074] Step S3: mapping the external heterogeneous data of the transport vehicle according to the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate a matrix of the external heterogeneous data of the transport vehicle;

[0075] Step S4: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle to generate the transport vehicle dynamic feature pattern data;

[0076] Step S5: Transmitting the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform vehicle identification feedback operation of the large-scale transport vehicle.

[0077] The radar technology of the present invention can work effectively in any weather conditions, including at night and in low visibility environments. This ensures that the monitoring of transport vehicles will not be interrupted regardless of environmental conditions, thereby improving the continuity and reliability of safety supervision. Radar signals can accurately measure the speed of transport vehicles and their relative positions to other objects, which is crucial to avoiding traffic congestion and accidents, especially on highways and in dense traffic environments. By analyzing radar signals, not only can the vehicle's motion state (such as speed and acceleration) data be obtained, but also the vehicle's static parameters (such as vehicle length and width) can be measured, providing a basis for further safety assessment and behavior analysis. High-resolution cameras provide clear visual images and can capture detailed features of vehicles, such as license plates, logos and vehicle colors, which are crucial to ensuring the accuracy and reliability of vehicle identification. Not only can the appearance features of the vehicle be extracted from the visual data, but the actual transportation behavior of the vehicle (such as loading status and loading and unloading process) can also be observed, which is conducive to more in-depth behavior analysis and risk assessment, and instant feedback on the status and behavior of the transport vehicle can be provided to support emergency response and immediate decision-making. Combining the static geometric data of the transporter (such as size and shape) with the appearance feature data (such as color and logo) to generate a comprehensive data matrix through heterogeneous data mapping, such fusion provides a comprehensive view of vehicle characteristics, which helps to more accurately identify and classify vehicles. Through detailed data level integration, richer information is provided to reduce misjudgment in the identification process, especially in the case of diverse vehicle types and models. The kinematic data (such as speed, acceleration) and transportation characteristics (such as loading status changes) of the transporter are analyzed to identify the behavior patterns of the vehicle. This is very useful for predicting vehicle behavior, assessing safety risks, and optimizing route planning. By monitoring the dynamic characteristics of the vehicle in real time, the system can instantly detect abnormal behavior, such as sudden acceleration or illegal lane changes, and take necessary safety measures. The timely transmission of integrated vehicle data to the terminal and the rapid execution of the identification and feedback process are crucial to the responsiveness and efficiency of the traffic management system. By using detailed vehicle characteristics and dynamic data, the management system can more effectively carry out traffic scheduling, route optimization and safety management, especially in complex situations involving large-scale transportation.

[0078] As an embodiment of the present invention, refer to Figure 1 The above is a schematic flow chart of the steps of a large-scale transport vehicle identification method based on radar and vision fusion of the present invention. In this embodiment, the large-scale transport vehicle identification method based on radar and vision fusion includes the following steps:

[0079] Step S1: Acquire the data of the supervision area on the way of the large-scale transport vehicle; collect the monitoring radar signal of the transport vehicle on the way through the radar monitoring equipment and the data of the supervision area on the way of the large-scale transport vehicle to generate the monitoring radar signal on the way of the transport vehicle; analyze the transport vehicle characteristics of the radar signal according to the monitoring radar signal on the way of the transport vehicle to generate the transport vehicle characteristic data, wherein the transport vehicle characteristic data includes the static geometry data of the transport vehicle and the kinematic data of the transport vehicle;

[0080] In an embodiment of the present invention, data of the supervision area on the way is extracted from the supervision system of large-scale transport vehicles, including transportation routes, monitoring points and environmental information. Using radar monitoring equipment deployed on the way of the transport vehicle, the radar signal of the transport vehicle is monitored in real time based on the supervision area data, and a radar signal for monitoring the transport vehicle on the way is generated, including a signal reflecting the static size and dynamic kinematic characteristics of the transport vehicle. The collected radar signal is processed and analyzed to extract the static geometric data of the transport vehicle, such as the length, width and height of the vehicle body, analyze the dynamic characteristics of the radar signal, and obtain the kinematic data of the transport vehicle, such as speed, acceleration, driving angle and trajectory.

[0081] Step S2: using high-resolution monitoring equipment and data from the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle; performing visual feature analysis on the way of the transport vehicle based on the visual data on the way of the transport vehicle, and generating visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static appearance feature data of the transport vehicle and transport feature data of the transport vehicle;

[0082] In an embodiment of the present invention, high-resolution monitoring equipment deployed in the supervision area is used to collect real-time video and images of the transport vehicles passing by. Based on the supervision area data, it is ensured that the monitoring equipment can fully cover the key parts of the transport vehicles and generate high-quality visual data. The collected visual data is processed to extract static transport vehicle appearance feature data, including license plate number, body color, vehicle logo and appearance shape. Through image processing technology, dynamic visual data is analyzed to extract the transport feature data of the transport vehicle, such as driving trajectory, turning angle and speed change.

[0083] Step S3: mapping the external heterogeneous data of the transport vehicle according to the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate a matrix of the external heterogeneous data of the transport vehicle;

[0084] In an embodiment of the present invention, key parameters such as length, width, height, etc. are extracted from the static geometric data of the transport vehicle. Important appearance features such as license plate number, body color, vehicle logo, and appearance shape are extracted from the static appearance feature data of the transport vehicle. The type and structure of the external information of the transport vehicle are analyzed, the association relationship between heterogeneous data is determined, and the architecture of the heterogeneous matrix is ​​designed, including the definition and configuration of data nodes and associated nodes. The static geometric data of the transport vehicle is mapped to the corresponding nodes of the heterogeneous matrix, and the static appearance feature data of the transport vehicle is mapped to the corresponding nodes of the heterogeneous matrix. The mapping data is integrated to generate a transport vehicle external heterogeneous data matrix to ensure that the data structure in the matrix is ​​complete and the association relationship is accurate.

[0085] Step S4: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle to generate the transport vehicle dynamic feature pattern data;

[0086] In an embodiment of the present invention, the kinematic data of the transport vehicle obtained by radar signal analysis is integrated with the transport characteristic data of the transport vehicle to form a complete dynamic data set. Key frame extraction is performed on the integrated dynamic data to identify representative moments or states in the transportation process, and key frame transport vehicle dynamic heterogeneous data is generated based on the key frame data. Time series analysis is performed on the key frame transport vehicle dynamic heterogeneous data to obtain the law of change of data in the time dimension, and time series key frame transport vehicle dynamic heterogeneous data is generated through time series analysis. Nonlinear spatiotemporal feature analysis is performed on the time series key frame transport vehicle dynamic heterogeneous data to extract the dynamic features of the transport vehicle in different time and space, and pattern recognition is performed based on the extracted dynamic features to generate the transport vehicle dynamic feature pattern data.

[0087] Step S5: Transmitting the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform vehicle identification feedback operation of the large-scale transport vehicle.

[0088] In the embodiment of the present invention, the generated external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle are packaged and processed, and the processed data is transmitted to the terminal system through a secure and reliable communication channel. In the terminal system, the transmitted data is received and parsed, and the vehicle identification operation of the large transport vehicle is performed according to the external heterogeneous data matrix and the dynamic characteristic pattern data, and the identification results and related feedback information are displayed to the user to assist supervision and decision-making.

[0089] Preferably, step S1 comprises the following steps:

[0090] Step S11: Obtaining data of the supervision area on the way of the large-scale transport vehicle;

[0091] Step S12: collecting radar signals for monitoring the transport vehicle en route through radar monitoring equipment and the data of the supervision area for the large-scale transport vehicle en route to generate radar signals for monitoring the transport vehicle en route, wherein the radar signals for monitoring the large-scale transport vehicle en route include static size monitoring radar signals and dynamic kinematic monitoring radar signals;

[0092] Step S13: extracting the amplitude peak value according to the static size monitoring radar signal to generate an amplitude peak size monitoring radar signal;

[0093] Step S14: analyzing static geometric data of the transport vehicle according to the amplitude peak size monitoring radar signal to generate static geometric data of the transport vehicle;

[0094] Step S15: Performing kinematic data analysis on the transport vehicle based on the dynamic kinematic monitoring radar signal to obtain the kinematic data of the transport vehicle.

[0095] By clearly setting the supervision area, the present invention can ensure the pertinence and efficiency of data collection, and provide accurate geographical and environmental background information for subsequent steps. The collection of radar signals not only covers the static size information of the vehicle, but also includes dynamic kinematic information, thereby realizing all-round monitoring of the transport vehicle. By extracting the amplitude peak of the radar signal, the key geometric features of the vehicle, such as vehicle length, vehicle width, etc., can be accurately identified, which are important parameters for evaluating vehicle size and category. Analysis of the amplitude peak size monitoring radar signal can generate detailed static geometric data of the vehicle and provide accurate information on the vehicle structure, which is key to compliance inspection and classification of transport vehicles. Based on the analysis of dynamic kinematic monitoring radar signals, kinematic data such as vehicle speed and acceleration can be obtained, which is very useful for evaluating vehicle driving behavior, safety, and adjusting traffic management strategies.

[0096] In an embodiment of the present invention, the supervision area data on the way is extracted from the supervision system of the large-scale transport vehicle, including the transportation route, the geographical location information of the monitoring point, the road conditions, etc., to ensure the real-time update of the supervision area data so as to accurately reflect the current position and driving environment of the transport vehicle. The radar monitoring equipment installed on the way of the transport vehicle receives the supervision area data to ensure that the radar equipment covers the entire supervision area. The radar equipment monitors the transport vehicle in real time, collects the radar signal on the way of the transport vehicle, and generates the transport vehicle monitoring radar signal. The transport vehicle monitoring radar signal includes two types of signals: static size monitoring radar signal: reflecting the static geometric dimensions of the transport vehicle, such as length, width and height, and is monitored by the radar equipment set in the station. Dynamic kinematic monitoring radar signal: reflecting the motion state of the transport vehicle, such as speed, acceleration and driving trajectory, and is monitored by the radar equipment set on the way. The static size monitoring radar signal is processed, the amplitude peak of the signal is extracted, and the amplitude peak size monitoring radar signal is generated, including: Fourier transforming the radar signal to identify the main frequency components of the signal and the corresponding amplitude peak. Based on the amplitude peak size monitoring radar signal, analyze the static geometric data of the transport vehicle, calculate the key dimensional parameters of the transport vehicle, such as body length, width, height, etc., and generate a detailed static geometric data report. Process the dynamic kinematic monitoring radar signal and analyze the kinematic data of the transport vehicle. Parse according to the receiving angle of the radar signal to generate signal receiving angle data. Through the signal receiving angle data, analyze the angle change of the transport vehicle and generate the transport vehicle angle data. Measure the receiving time difference of the radar signal to generate signal receiving time difference data. Combine the signal receiving time difference data and the signal receiving angle data to evaluate the path of the transport vehicle and generate a transport vehicle path data evaluation report. Analyze the frequency change of the radar signal to generate signal frequency change data. Combine the signal frequency change data, signal receiving time difference data and signal receiving angle data to analyze the Doppler effect, calculate the travel speed of the transport vehicle, and generate the transport vehicle speed data. Integrate the transport vehicle angle data, transport vehicle path data evaluation data and transport vehicle speed data to generate complete transport vehicle kinematic data.

[0097] Preferably, step S15 comprises the following steps:

[0098] Step S151: performing signal receiving angle analysis according to the dynamic kinematics monitoring radar signal to generate signal receiving angle data, and performing transport vehicle angle analysis through the signal receiving angle data to generate transport vehicle angle data;

[0099] Step S152: performing signal reception time difference measurement and analysis according to the dynamic kinematics monitoring radar signal to generate signal reception time difference data, and performing transport vehicle path data evaluation processing through the signal reception time difference data and the signal reception angle data to generate transport vehicle path data evaluation data;

[0100] Step S153: performing signal frequency change analysis according to the dynamic kinematics monitoring radar signal to generate signal frequency change analysis data, and performing Doppler effect transport vehicle speed analysis through the signal frequency change analysis data, signal reception time difference data, and signal reception angle data to generate transport vehicle speed data;

[0101] Step S154: Integrate the transport vehicle angle data, the transport vehicle path data evaluation data and the transport vehicle speed data to generate the transport vehicle kinematic data.

[0102] The present invention accurately measures and determines the direction and angle of the transport vehicle by analyzing the received signal angle, which is critical for understanding the vehicle's driving direction and its specific position on the road. By measuring the time difference of signal reception, the distance between the transport vehicle and the radar device is accurately evaluated. This provides accurate vehicle location information for traffic management, helps prevent collisions and optimize traffic flow. By analyzing the changes in signal frequency, the speed of the transport vehicle can be accurately measured using the Doppler effect. This is crucial for monitoring the vehicle's driving speed, predicting traffic flow changes, and enforcing speed limit policies. The transport vehicle's angle data, path data, and speed data are integrated to generate a comprehensive kinematic data set. This data integration provides a panoramic view of the transport vehicle's behavior, enhances the understanding of its dynamic performance, and supports more accurate behavior prediction and safety analysis.

[0103] In an embodiment of the present invention, a radar monitoring device is used to receive the radar signal reflected by the transport vehicle, and the radar signal is monitored based on dynamic kinematics to determine the receiving angle of the radar signal, and the received radar signal is parsed to calculate the angle at which the signal reaches the sensor, and triangulation or other appropriate geometric calculation methods are used to generate signal receiving angle data. By analyzing the signal receiving angle data, the angle change of the transport vehicle is determined, and the angle change information is organized into a data report to generate the transport vehicle angle data. Based on the dynamic kinematics monitoring radar signal, the time when the radar signal is sent and received is recorded, and the time difference from the transmission to the reception of the radar signal is measured to generate the signal receiving time difference data. The signal receiving time difference data is analyzed using an algorithm to determine the relative position change of the transport vehicle, and the path evaluation is performed in combination with the signal receiving angle data. The driving path and direction of the transport vehicle are evaluated through the signal receiving time difference data and the signal receiving angle data, and the evaluation results are integrated to generate the transport vehicle path data. According to the dynamic kinematics monitoring radar signal, the frequency change of the radar signal is recorded, the frequency change of the signal is analyzed, and the signal frequency change parsing data is generated, and the relative speed change of the transport vehicle is calculated using the Doppler effect principle. Combine the signal frequency change analysis data, signal reception time difference data and signal reception angle data to analyze the speed of the transport vehicle, generate the transport vehicle speed data, and provide the speed change information of the transport vehicle. Collect the transport vehicle angle data, transport vehicle path data evaluation data and transport vehicle speed data, integrate the angle data, path data and speed data to ensure the consistency and integrity of the data, use the data integration algorithm to merge different types of data into a comprehensive data set, and generate complete transport vehicle kinematic data.

[0104] Preferably, step S2 comprises the following steps:

[0105] Step S21: using high-resolution monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle, wherein the visual data on the way of the transport vehicle includes static transport vehicle image data and dynamic transport vehicle image data;

[0106] Step S22: extracting the appearance features of the static transport vehicle according to the static transport vehicle image data to generate the appearance feature data of the static transport vehicle;

[0107] Step S23: Analyze the transport characteristics of the transport vehicle based on the dynamic transport vehicle image data to generate transport characteristic data of the transport vehicle.

[0108] The present invention uses high-resolution monitoring equipment to obtain visual data of the transport vehicle, including static and dynamic images. This high-quality image acquisition provides detailed visual evidence for subsequent in-depth analysis, enhancing the availability of data and the depth of analysis. The data collected in the supervision area on the way of the large-scale transport vehicle not only records the behavior of the vehicle, but also captures the situation of the surrounding environment, which helps to more comprehensively evaluate the operating environment of the vehicle and possible interactions or conflicts. The appearance features of the transport vehicle, such as model, color, license plate, etc., are extracted from the static image. These detailed appearance feature data are crucial for the confirmation and classification of vehicle identity, especially in situations where different vehicles or vehicle types need to be accurately identified. By accurately extracting the static features of each vehicle, the error in the vehicle identification process is reduced, and the reliability and accuracy of the overall monitoring system are improved. Based on the dynamic image data of the transport vehicle, analyzing its transportation characteristics, such as the vehicle's driving speed, driving situation, load changes, etc., is extremely important for understanding the behavior pattern of the vehicle during actual transportation. Through in-depth analysis of the transportation characteristics of the transport vehicle, potential safety issues or efficiency bottlenecks can be identified, and measures can be taken to optimize the transportation route or improve the safety of transportation operations.

[0109] In an embodiment of the present invention, high-resolution monitoring equipment is deployed at key locations on the way of a large-scale transport vehicle to ensure that the monitoring equipment covers the main sections and regulatory areas on the way of the transport vehicle, and monitors the driving conditions of the transport vehicle in real time. Using high-resolution monitoring equipment, according to the regulatory area data on the way of the large-scale transport vehicle, real-time video and image acquisition is performed to generate visual data on the way of the transport vehicle, including static transport vehicle image data and dynamic transport vehicle image data. Capture a clear image of the transport vehicle at a specific time point. Record the continuous motion image of the transport vehicle to reflect the driving trajectory and dynamic changes of the transport vehicle. The collected static transport vehicle image data is preprocessed, including image denoising, enhancement and correction operations to improve the image quality, and the image segmentation technology is used to extract the main part of the transport vehicle and remove background interference. Based on the preprocessed static image, the appearance features of the transport vehicle are extracted, including: using optical character recognition (OCR) technology to extract the license plate number; determining the body color through color histogram and color space conversion; using template matching or deep learning methods to identify the vehicle logo; extracting the appearance shape features of the transport vehicle through edge detection and shape descriptors. Integrate the extracted appearance feature information to generate static transport vehicle appearance feature data. The collected dynamic transport vehicle image data is preprocessed, including video denoising, frame rate adjustment, and image stabilization, to segment the continuous transport vehicle motion images and ensure that each frame is clear and blur-free. The dynamic image is binarized and converted into a black and white binary image to highlight the outline and motion trajectory of the transport vehicle. The image segmentation algorithm is used to segment the binarized dynamic image into several sub-areas to extract the motion area of ​​the transport vehicle. The segmented dynamic image is subjected to motion blur analysis to extract motion blur distortion parameters to describe the degree of blur of the transport vehicle during movement. Based on the extracted motion blur distortion parameters, the dynamic image is motion-blur corrected to generate a clear dynamic image. The transportation feature data is extracted from the corrected dynamic image, including information such as driving speed, acceleration, turning angle, and driving path. The extracted transportation feature information is integrated to generate complete transportation feature data of the transport vehicle.

[0110] Preferably, step S23 includes the following steps:

[0111] Step S231: performing image binarization processing on the dynamic transport vehicle image data to generate binary dynamic transport vehicle image data;

[0112] Step S232: performing image segmentation according to the binary dynamic transport vehicle image data to generate segmented dynamic transport vehicle image data;

[0113] Step S233: analyzing the motion blur distortion parameters of the transport vehicle according to the segmented dynamic transport vehicle image data to generate motion blur distortion parameters;

[0114] Step S234: performing motion blur distortion correction processing on the dynamic transport vehicle image data according to the motion blur distortion parameters to generate corrected dynamic transport vehicle image data;

[0115] Step S235: Performing transportation characteristic analysis based on the corrected dynamic transport vehicle image data to generate transport characteristic data of the transport vehicle.

[0116] The present invention simplifies the complexity of the image by converting the dynamic transport vehicle image data into a binary format, making the subsequent processing steps more efficient and accurate, especially in image segmentation and feature extraction. The segmentation of the binary image makes it easier to accurately identify the outline and other important parts of the transport vehicle, which helps to effectively separate the vehicle from the complex background and improve the accuracy of target detection. Accurately analyzing the blur distortion parameters caused by vehicle motion provides necessary information for subsequent image correction, so that the image quality can be improved in a targeted manner. Applying motion blur distortion parameters for image correction effectively reduces or eliminates the blur caused by the rapid movement of the vehicle, restores the clarity of the image, and helps to perform subsequent feature analysis more accurately. The corrected image data provides clearer visual information, making the transportation feature analysis of the transport vehicle more accurate, such as vehicle load status, loading and unloading activities and behavior patterns, etc., which are crucial for transportation supervision and safety analysis.

[0117] In the embodiment of the present invention, the dynamic transport vehicle image data is subjected to denoising, the noise in the image is removed, the image is enhanced, the contrast and brightness of the image are improved, the outline of the transport vehicle is ensured to be clearly visible, and the pre-processed image is converted into a grayscale image. An appropriate binarization threshold is selected to convert the grayscale image into a binary image, generate binary dynamic transport vehicle image data, and ensure that the outline of the transport vehicle is clearly presented in the binary image. Morphological operations such as corrosion and expansion are performed on the binary dynamic transport vehicle image data to remove noise points and fill holes, further enhance the outline of the transport vehicle, use an image segmentation algorithm based on connected components, segment the binary image into several regions, identify the outline of the transport vehicle, mark each segmented region, generate segmented dynamic transport vehicle image data, and ensure that each region represents a part of the transport vehicle. Analyze and segment the dynamic transport vehicle image data to identify areas in the image where motion blur may exist. Use image gradient analysis technology (such as gradient-based edge monitoring algorithm) to detect blur edges and blur levels in the image. Based on the analysis results of the blur area, extract motion blur distortion parameters, including blur direction, blur level, blur kernel size and other parameters, and generate a detailed motion blur distortion parameter report. According to the motion blur distortion parameters, select an appropriate motion blur correction algorithm, such as Wiener filtering, blind deconvolution, etc., apply the blur correction algorithm to the dynamic transport vehicle image data, correct the motion blur in the image, restore the clarity of the image, ensure that the details of the corrected image are retained as much as possible, and generate corrected dynamic transport vehicle image data. Extract the key feature points of the transport vehicle from the corrected dynamic transport vehicle image data, such as the front, rear, tire position, etc., and use feature detection algorithms (such as SIFT, SURF, ORB, etc.) to identify and describe these feature points. Track the motion trajectory of the transport vehicle in the image sequence, use the optical flow method or other motion tracking algorithms to calculate the displacement and motion direction of the transport vehicle in the continuous frames, and extract preliminary transport features, including the transport vehicle's speed, acceleration, turning angle, driving path and contour changes, etc. Based on the position data in the image sequence, establish the vehicle's driving trajectory model, use polynomial fitting or spline curve fitting methods to model the transport vehicle's trajectory, and extract trajectory features, such as turning angle, turning radius, straight-line distance, etc. Integrate the preliminary transportation features and trajectory features to obtain the transport vehicle's transportation feature data.

[0118] Preferably, step S233 includes the following steps:

[0119] The middle nodes of the segmented dynamic transport vehicle image data are selected as the dynamic transport vehicle mass point parameters, and the dynamic transport vehicle mass point parameters are analyzed in time series to generate the time series dynamic transport vehicle mass point parameters;

[0120] Performing fuzzy area recognition on the segmented dynamic transport vehicle image data to generate dynamic transport vehicle fuzzy area data;

[0121] According to the time series dynamic transport vehicle particle parameters and the dynamic transport vehicle blur area data, the motion blur distortion parameters of the transport vehicle are analyzed to generate motion blur distortion parameters.

[0122] The present invention selects the middle nodes in the image data as particles and performs time series analysis on them, which can accurately track the movement path and speed of the transport vehicle, help better understand the specific movement state of the vehicle during monitoring, and provide key data for the motion model. The segmented dynamic transport vehicle image data is subjected to fuzzy area identification to clarify the key areas of image processing, that is, those fuzzy areas caused by rapid movement, which helps to make subsequent image correction more targeted and improve processing efficiency and effect. Combined with the time series dynamic transport vehicle particle parameters and the dynamic transport vehicle fuzzy area data, motion blur distortion parameter analysis is performed, which can accurately identify the specific blur type and degree caused by the rapid movement of the vehicle, so that the subsequent image correction work can be carried out based on the actual distortion situation, greatly improving the image availability and analysis accuracy.

[0123] In the embodiment of the present invention, from the segmented dynamic transport vehicle image data, the middle node of each segmented area is selected as the particle of the dynamic transport vehicle. The middle node is usually the centroid of the segmented area, representing the center position of the area. The spatial coordinates of each middle node are recorded as particle parameters, and the particle parameters at each moment are recorded to form a series of time series data. The time series data of the particle parameters are analyzed to identify the trajectory and motion law of the particle over time, and the time series analysis method, such as Kalman filtering, particle filtering, etc., is used to smooth and predict the particle motion trajectory, generate the time series dynamic transport vehicle particle parameters, and describe the change trend of the particle in the time dimension. The segmented dynamic transport vehicle image data is subjected to fuzzy area detection, and the part with motion blur in the image is identified. The image processing techniques such as gradient detection and Laplace transform are used to detect the fuzzy edge in the image. The detected fuzzy area is marked, the boundary and area of ​​the fuzzy area are recorded, and the dynamic transport vehicle fuzzy area data is generated to describe the detailed information of each fuzzy area in the image. The particle parameters of the time series dynamic transport vehicle and the blurred area data of the dynamic transport vehicle are fused to ensure the consistency and correlation of the particle parameters and the blurred area data, forming a complete data set. Based on the fused data, the motion blur distortion of the transport vehicle is analyzed in detail. Based on the particle parameters of the time series dynamic transport vehicle, the motion trajectory model of the transport vehicle particle is established. The motion trajectory of the particle is fitted using polynomial fitting or spline curve fitting methods. The motion characteristics of the particle are extracted from the motion trajectory, including speed change, acceleration change and motion direction change, and these characteristics are correlated with the blurred area data for analysis. Assuming that the blur kernel is a linear motion blur kernel, the shape of the blur kernel is preliminarily estimated according to the main direction of the blurred area and the motion direction of the particle. The shape of the blur kernel should be consistent with the motion direction of the particle and the gradient direction of the blurred area. According to the blur degree and the motion speed of the particle, the size (length and width) of the blur kernel is preliminarily estimated. The size of the blur kernel should be proportional to the motion speed of the particle and the time interval. The initial blur kernel is generated. The shape and size are determined according to the preliminary estimation results of the previous step. The initial blur kernel is applied to the blurred area to simulate the initial blur effect. Use the deconvolution algorithm to deblur the blurred area, gradually adjust the shape and size of the blur kernel, and continuously optimize the blur kernel to best match the actual blur effect. Use iterative optimization algorithms (such as gradient descent) to further adjust the parameters (direction, size, shape) of the blur kernel. In each iteration, evaluate the deblurring effect and adjust the blur kernel parameters based on the evaluation results. Extract blur distortion parameters, including blur direction, blur degree, and blur kernel size, organize and summarize the extracted motion blur distortion parameters, and generate a detailed parameter analysis report to ensure that the report contains the distortion parameters of each blurred area and provides accurate blur distortion information.

[0124] Preferably, step S3 comprises the following steps:

[0125] Step S31: extracting the external information type of the transport vehicle from the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate the external information type of the transport vehicle;

[0126] Step S32: performing heterogeneous matrix architecture analysis according to the external information type of the transport vehicle to construct a heterogeneous matrix architecture;

[0127] Step S33: transmitting the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to the heterogeneous matrix architecture for data mapping of the external heterogeneous data of the transport vehicle to generate an external heterogeneous data matrix of the transport vehicle.

[0128] The present invention can more effectively classify and standardize this information by extracting key information types from the static geometric data and static appearance feature data of the transport vehicle, which not only helps to organize and manage the data, but also improves the efficiency and accuracy of subsequent data processing steps. Heterogeneous matrix architecture analysis is performed based on the extracted external information types of the transport vehicle, aiming to build a data architecture that can effectively process and store this information, optimize data query, storage and access efficiency, and provide a basis for more complex data analysis and decision support. The static geometric data and appearance feature data of the transport vehicle are mapped to the constructed heterogeneous matrix architecture to generate a transport vehicle external heterogeneous data matrix, which can provide a full view of vehicle characteristics and greatly enhance the application value of the data, especially when performing complex vehicle identification and classification tasks.

[0129] In the embodiment of the present invention, key parameters are extracted from the static geometric data of the transport vehicle obtained in the previous step, including geometric dimension data such as the length, width, height, and axle spacing of the transport vehicle. Important features are extracted from the static appearance feature data of the transport vehicle obtained in the previous step, including appearance features such as license plate number, body color, car logo, and body shape. Combined with the static geometric data and appearance feature data, the external information type of the transport vehicle is generated, and each information type includes corresponding data fields and descriptions, such as "license plate number-character type", "body color-color value", etc., to generate a detailed list of external information types of the transport vehicle, including all extracted geometric and appearance feature data. According to the generated external information type of the transport vehicle, the information type is analyzed and classified, and the information type is divided into two categories: geometric information and appearance feature information. A heterogeneous matrix architecture is designed to determine the position and representation of each information type in the matrix, define the nodes of the heterogeneous matrix, each node represents an information type, determine the association relationship between the nodes, and design the connection method of the heterogeneous matrix. According to the designed architecture, a heterogeneous matrix is ​​constructed to ensure that the matrix structure is clear and the association relationship is clear, providing a basis for subsequent data mapping. Collect and organize the static geometric data and static appearance feature data of the transport vehicle, map the static geometric data and static appearance feature data to the corresponding nodes of the heterogeneous matrix according to the information type, fill in the data for each node to ensure the accuracy and consistency of the node data, integrate the data of all nodes, and generate a complete transport vehicle external heterogeneous data matrix to ensure that the heterogeneous matrix contains all necessary geometric and appearance feature data, and the relationship between the data is clear and unambiguous.

[0130] Preferably, step S32 includes the following steps:

[0131] Perform unique identification external information type analysis based on the external information type of the transport vehicle to generate unique identification external information type data;

[0132] Designing unique identification nodes of heterogeneous matrices by uniquely identifying external information type data, and designing heterogeneous matrix nodes by transport vehicle external information type;

[0133] Perform heterogeneous matrix node association processing on the heterogeneous matrix unique identification node and the heterogeneous matrix node to generate heterogeneous matrix node association data;

[0134] A heterogeneous matrix architecture is established for the unique identification nodes of the heterogeneous matrix and the heterogeneous matrix nodes according to the heterogeneous matrix node association data to construct a heterogeneous matrix architecture.

[0135] The present invention ensures that each type of data has a clear identifier by performing unique identification analysis on the external information type of the transport vehicle, which helps to avoid data duplication and errors and increase the overall accuracy and reliability of the database. The unique identification node in the heterogeneous matrix is ​​designed using the unique identification external information type data, which improves the logic and search efficiency of the data structure and makes data retrieval faster and more accurate. By associating the unique identification node and other nodes of the heterogeneous matrix, a complex data relationship network is created, which not only helps to understand the relationship between the data, but also improves the depth and breadth of data analysis. The heterogeneous matrix architecture is constructed based on the node association data, providing a highly flexible and scalable platform that supports various types of data analysis and processing, not only meeting current needs, but also easily adapting to future expansion and modification.

[0136] In the embodiment of the present invention, according to the type of external information of the transport vehicle extracted, the information type is divided into different categories, such as static geometric information and appearance feature information, and a unique identification information type is selected from each information category, such as the license plate number as the unique identification of the appearance feature information, and the vehicle ID as the unique identification of the static geometric information, to ensure that the selected unique identification can uniquely identify each transport vehicle, avoid duplication and conflict, generate unique identification external information type data, including information such as the license plate number and the vehicle ID, to ensure that these data are unique and representative in the system, and based on the unique identification external information type data, design a unique identification node of a heterogeneous matrix, each unique identification node represents a unique identification information type, such as a license plate number node and a vehicle ID node, and design other heterogeneous matrix nodes according to the type of external information of the transport vehicle, each node corresponds to a specific information type, such as a body color node, a vehicle logo node, a body length node, etc. Define attributes and data fields for each node, including node name, data type, data source, etc., to ensure that the attributes and data fields of each node are clear and complete. According to the logical relationship between information types, define the association rules of heterogeneous matrix nodes, determine which nodes are directly associated with each other, and which nodes need to be associated through unique identification nodes, perform association processing on the unique identification nodes and ordinary nodes of the heterogeneous matrix, and generate heterogeneous matrix node association data. Ensure that the association relationship of each node is clear and logically reasonable, verify the accuracy of the node association data, ensure that the association relationship is correct, test and adjust the node association data, and ensure that the data is complete and consistent. Based on the node association data, design the overall architecture of the heterogeneous matrix, confirm the position and connection method of each node in the matrix, and actually build the designed architecture through programming or tools, configure each node according to the designed position and association method, and establish a complete heterogeneous matrix architecture.

[0137] Preferably, step S4 comprises the following steps:

[0138] Step S41: integrating the transport vehicle kinematic data and the transport vehicle transport characteristic data into dynamic heterogeneous transport vehicle data;

[0139] Step S42: performing dynamic key frame analysis based on the dynamic heterogeneous data of the transport vehicle to generate key frame dynamic heterogeneous data of the transport vehicle;

[0140] Step S43: performing time series analysis on the key frame transport vehicle dynamic heterogeneous data to generate time series key frame transport vehicle dynamic heterogeneous data, and performing nonlinear transport vehicle dynamic spatiotemporal feature analysis based on the time series key frame transport vehicle dynamic heterogeneous data to generate transport vehicle dynamic spatiotemporal feature data;

[0141] Step S44: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle dynamic spatiotemporal feature data to generate transport vehicle dynamic feature pattern data.

[0142] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed implementation step flow chart of step S4 in the embodiment, step S4 in this embodiment includes:

[0143] Step S41: integrating the transport vehicle kinematic data and the transport vehicle transport characteristic data into dynamic heterogeneous transport vehicle data;

[0144] In the embodiment of the present invention, the transport vehicle kinematic data and transport vehicle transport characteristic data generated in the previous steps are collected to ensure the integrity and accuracy of the data and prepare a data set for integration. The transport vehicle kinematic data and the transport vehicle transport characteristic data are converted into a unified data format to ensure that all data fields and data types are consistent, which is convenient for subsequent integration processing. The transport vehicle kinematic data and the transport vehicle transport characteristic data are integrated to form a comprehensive transport vehicle dynamic heterogeneous data set to ensure seamless integration of data and avoid duplication and data loss.

[0145] Step S42: performing dynamic key frame analysis based on the dynamic heterogeneous data of the transport vehicle to generate key frame dynamic heterogeneous data of the transport vehicle;

[0146] In an embodiment of the present invention, based on the dynamic heterogeneous data of the transport vehicle, an algorithm is used to extract key frames in the transport vehicle's motion process, including rate-of-change methods (such as speed change rate, acceleration change rate), event-based methods (such as turning, sudden braking), etc. The key frames represent the important states and changes of the transport vehicle at different time points, such as turning, acceleration, deceleration, etc. The rate of change is calculated for the kinematic data to identify the time points when the changes are obvious. For example, by calculating the rate of change of speed and acceleration, the moment when the transport vehicle changes significantly during driving is found. An event-based method is used to detect key events (such as turning, sudden braking, acceleration, etc.) during driving of the transport vehicle. These events are usually important moments in the driving process. Combined with the rate of change and event detection results, the key frames representing the driving process of the transport vehicle are screened out to ensure that the key frames can fully describe the driving behavior and dynamic characteristics of the transport vehicle. The annotated key frame data is integrated with the original dynamic heterogeneous data of the transport vehicle to form a complete key frame data set.

[0147] Step S43: performing time series analysis on the key frame transport vehicle dynamic heterogeneous data to generate time series key frame transport vehicle dynamic heterogeneous data, and performing nonlinear transport vehicle dynamic spatiotemporal feature analysis based on the time series key frame transport vehicle dynamic heterogeneous data to generate transport vehicle dynamic spatiotemporal feature data;

[0148] In an embodiment of the present invention, the dynamic heterogeneous data of the key frame transport vehicle generated in the previous steps are collected to ensure the integrity and consistency of the data, prepare a data set for time series analysis, arrange the key frame data in chronological order, and construct a time series. Each key frame contains information such as timestamp, kinematic data, and transport characteristic data. According to the characteristics of the key frame data, a suitable time series analysis method is selected, such as dynamic time warping (DTW), moving average (MA), autoregressive integrated moving average model (ARIMA), etc., and the selected time series analysis method is applied to extract the time pattern in the key frame data, analyze the time change law of the data, and identify the behavior characteristics of the transport vehicle in different time periods. The extracted time pattern and the analysis results are integrated to generate the dynamic heterogeneous data of the time series key frame transport vehicle, ensuring that the time series data accurately reflects the time change and dynamic characteristics of the key frame. According to the characteristics of the dynamic heterogeneous data of the time series key frame transport vehicle, select the appropriate nonlinear analysis method. Common nonlinear analysis methods include nonlinear dynamic analysis, fractal analysis, etc. Preprocess the time series key frame data to remove noise and outliers, normalize the data to ensure that data of different dimensions have the same scale, apply the selected nonlinear analysis method, extract the dynamic spatiotemporal characteristics of the transport vehicle, analyze the nonlinear relationship in the data, and identify complex spatiotemporal behavior patterns, such as periodic changes, chaotic phenomena, etc. The extracted nonlinear features are fused with the original time series data to form complete dynamic spatiotemporal feature data, ensuring the multi-dimensional and multi-scale characteristics of the feature data and providing rich behavior descriptions. Integrate the nonlinear analysis results to generate dynamic spatiotemporal feature data of the transport vehicle.

[0149] Step S44: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle dynamic spatiotemporal feature data to generate transport vehicle dynamic feature pattern data.

[0150] In the embodiment of the present invention, the dynamic spatiotemporal feature data of the transport vehicle is preprocessed, outliers and missing values ​​are cleared, and the feature data is normalized to ensure that different feature data have the same scale. According to the characteristics of the feature data and the recognition requirements, a suitable pattern recognition algorithm is selected. Commonly used algorithms include support vector machine (SVM), random forest (RF), convolutional neural network (CNN), long short-term memory network (LSTM), etc. Considering the complexity, computational efficiency and recognition accuracy of the algorithm, the optimal recognition algorithm is selected. The preprocessed feature data is divided into a training set and a test set, usually divided in a ratio of 7:3 or 8:2, to ensure the representativeness and diversity of the training set and test set data and avoid data bias. Use the training set data to train the selected pattern recognition model, perform multiple rounds of training and parameter adjustment, optimize the performance and recognition accuracy of the model, and use the cross-validation method to evaluate the generalization ability and stability of the model. Use the test set data to verify the trained model, evaluate the model's recognition accuracy, recall rate, F1 value and other performance indicators, and further adjust and optimize the model parameters according to the verification results to improve the recognition effect of the model. The dynamic spatiotemporal feature data of the transport vehicle is input into the trained pattern recognition model to perform feature pattern recognition, identify the dynamic feature patterns of the transport vehicle, such as different driving states, behavior patterns, etc., organize and summarize the identified feature patterns, and generate the dynamic feature pattern data of the transport vehicle.

[0151] The present invention integrates the kinematic data and transportation characteristic data of the transport vehicle into a unified data set, providing a comprehensive perspective on the dynamic behavior of the transport vehicle, and increasing the consistency of data processing and the depth of analysis. By analyzing and determining the key frames in the dynamic data, important moments and turning points in the transportation process are identified, which is crucial to understanding the behavior of the transport vehicle and making corresponding responses. Time series analysis is performed on the key frame dynamic heterogeneous data, and further nonlinear dynamic spatiotemporal characteristic analysis is performed to reveal the behavior patterns and dynamic changes of the transport vehicle, which are very valuable for predicting future behavior and conducting risk assessments. Based on the dynamic spatiotemporal characteristic data of the transport vehicle, dynamic characteristic patterns are identified to effectively predict the possible behavioral trends and risks of the transport vehicle, support decision-making and emergency response plans, and thus improve transportation safety and efficiency.

[0152] This specification also provides a large-scale transport vehicle identification system based on radar and vision fusion, which is used to execute the large-scale transport vehicle identification method based on radar and vision fusion as described above. The large-scale transport vehicle identification system based on radar and vision fusion includes:

[0153] The radar signal transport vehicle characteristic analysis module is used to obtain the data of the supervision area on the way of the large-scale transport vehicle; collect the transport vehicle monitoring radar signal on the way through the radar monitoring equipment and the data of the supervision area on the way of the large-scale transport vehicle to generate the transport vehicle monitoring radar signal on the way; analyze the transport vehicle characteristics of the radar signal according to the transport vehicle monitoring radar signal on the way to generate the transport vehicle characteristic data, wherein the transport vehicle characteristic data includes the transport vehicle static geometry data and the transport vehicle kinematic data;

[0154] The transport vehicle visual feature analysis module is used to collect visual data of the transport vehicle on the way by using high-resolution monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle, and generate visual data of the transport vehicle on the way; perform visual feature analysis on the transport vehicle on the way according to the visual data on the way of the transport vehicle, and generate visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static transport vehicle appearance feature data and transport feature data of the transport vehicle;

[0155] A transport vehicle external heterogeneous data mapping module is used to map the transport vehicle external heterogeneous data according to the transport vehicle static geometric data and static transport vehicle appearance feature data to generate a transport vehicle external heterogeneous data matrix;

[0156] The transport vehicle dynamic feature pattern recognition module is used to perform transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle, and generate the transport vehicle dynamic feature pattern data;

[0157] The transport vehicle identification feedback module is used to transmit the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform the vehicle identification feedback operation of the large-scale transport vehicle.

[0158] The beneficial effect of the present application is that the method for identifying large-scale transport vehicles based on radar and vision fusion of the present invention utilizes the ability of radar to provide accurate distance and speed data under any weather and lighting conditions, and combines the detailed image and appearance information provided by the visual system to ensure high-precision vehicle identification and classification under various environmental conditions. Through in-depth analysis of kinematic data and transportation characteristics, the identification method can not only capture the vehicle's real-time position and status, but also predict the risk of large-scale transport vehicles through timing and dynamic feature analysis, providing support for traffic safety and route planning. And through the fusion analysis of radar and vision, more identification information of large-scale transport vehicles is provided.

[0159] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0160] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for identifying large-scale transport vehicles based on radar and vision fusion, characterized in that: The following steps are involved: Step S1: Obtaining data on the supervision area where the large-scale transport vehicle is en route; Collecting radar signals for monitoring the transport vehicle on the way through radar monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle to generate radar signals for monitoring the transport vehicle on the way; analyzing the transport vehicle characteristics of the radar signals based on the radar signals for monitoring the transport vehicle on the way to generate transport vehicle characteristic data, wherein the transport vehicle characteristic data includes static geometric data of the transport vehicle and kinematic data of the transport vehicle; Wherein, step S1 comprises the following steps: Step S11: Obtaining data of the supervision area on the way of the large-scale transport vehicle; Step S12: collecting radar signals for monitoring the transport vehicle en route through radar monitoring equipment and the data of the supervision area for the large-scale transport vehicle en route to generate radar signals for monitoring the transport vehicle en route, wherein the radar signals for monitoring the large-scale transport vehicle en route include static size monitoring radar signals and dynamic kinematic monitoring radar signals; Step S13: extracting the amplitude peak value according to the static size monitoring radar signal to generate an amplitude peak size monitoring radar signal; Step S14: analyzing static geometric data of the transport vehicle according to the amplitude peak size monitoring radar signal to generate static geometric data of the transport vehicle; Step S15: performing kinematic data analysis of the transport vehicle based on the dynamic kinematic monitoring radar signal to obtain kinematic data of the transport vehicle; Wherein, step S15 comprises the following steps: Step S151: performing signal receiving angle analysis according to the dynamic kinematics monitoring radar signal to generate signal receiving angle data, and performing transport vehicle angle analysis through the signal receiving angle data to generate transport vehicle angle data; Step S152: performing signal reception time difference measurement and analysis according to the dynamic kinematics monitoring radar signal to generate signal reception time difference data, and performing transport vehicle path data evaluation processing through the signal reception time difference data and the signal reception angle data to generate transport vehicle path data evaluation data; Step S153: performing signal frequency change analysis according to the dynamic kinematics monitoring radar signal to generate signal frequency change analysis data, and performing Doppler effect transport vehicle speed analysis through the signal frequency change analysis data, signal reception time difference data, and signal reception angle data to generate transport vehicle speed data; Step S154: integrating the transport vehicle angle data, the transport vehicle path data evaluation data and the transport vehicle speed data to generate the transport vehicle kinematic data; Step S2: using high-resolution monitoring equipment and data from the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle; performing visual feature analysis on the way of the transport vehicle based on the visual data on the way of the transport vehicle, and generating visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static appearance feature data of the transport vehicle and transport feature data of the transport vehicle; Step S3: mapping the external heterogeneous data of the transport vehicle according to the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate a matrix of the external heterogeneous data of the transport vehicle; Step S4: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle to generate the transport vehicle dynamic feature pattern data; Step S5: Transmitting the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform vehicle identification feedback operation of the large-scale transport vehicle.

2. The large-scale transport vehicle identification method based on radar and vision fusion according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: using high-resolution monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle to collect visual data on the way of the transport vehicle, and generating visual data on the way of the transport vehicle, wherein the visual data on the way of the transport vehicle includes static transport vehicle image data and dynamic transport vehicle image data; Step S22: extracting the appearance features of the static transport vehicle according to the static transport vehicle image data to generate the appearance feature data of the static transport vehicle; Step S23: Analyze the transport characteristics of the transport vehicle based on the dynamic transport vehicle image data to generate transport characteristic data of the transport vehicle.

3. The large-scale transport vehicle identification method based on radar and vision fusion according to claim 2 is characterized in that: Step S23 includes the following steps: Step S231: performing image binarization processing on the dynamic transport vehicle image data to generate binary dynamic transport vehicle image data; Step S232: performing image segmentation according to the binary dynamic transport vehicle image data to generate segmented dynamic transport vehicle image data; Step S233: analyzing the motion blur distortion parameters of the transport vehicle according to the segmented dynamic transport vehicle image data to generate motion blur distortion parameters; Step S234: performing motion blur distortion correction processing on the dynamic transport vehicle image data according to the motion blur distortion parameters to generate corrected dynamic transport vehicle image data; Step S235: Performing transportation characteristic analysis based on the corrected dynamic transport vehicle image data to generate transport characteristic data of the transport vehicle.

4. The large-scale transport vehicle identification method based on radar and vision fusion according to claim 3 is characterized in that: Step S233 includes the following steps: The middle nodes of the segmented dynamic transport vehicle image data are selected as the dynamic transport vehicle mass point parameters, and the dynamic transport vehicle mass point parameters are analyzed in time series to generate the time series dynamic transport vehicle mass point parameters; Performing fuzzy area recognition on the segmented dynamic transport vehicle image data to generate dynamic transport vehicle fuzzy area data; According to the time series dynamic transport vehicle particle parameters and the dynamic transport vehicle blur area data, the motion blur distortion parameters of the transport vehicle are analyzed to generate motion blur distortion parameters.

5. The large-scale transport vehicle identification method based on radar and vision fusion according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: extracting the external information type of the transport vehicle from the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to generate the external information type of the transport vehicle; Step S32: performing heterogeneous matrix architecture analysis according to the external information type of the transport vehicle to construct a heterogeneous matrix architecture; Step S33: transmitting the static geometric data of the transport vehicle and the static appearance feature data of the transport vehicle to the heterogeneous matrix architecture for data mapping of the external heterogeneous data of the transport vehicle to generate an external heterogeneous data matrix of the transport vehicle.

6. The method for identifying large-scale transport vehicles based on radar and vision fusion according to claim 5 is characterized in that: Step S32 includes the following steps: Perform unique identification external information type analysis based on the external information type of the transport vehicle to generate unique identification external information type data; Designing unique identification nodes of heterogeneous matrices by uniquely identifying external information type data, and designing heterogeneous matrix nodes by transport vehicle external information type; Perform heterogeneous matrix node association processing on the heterogeneous matrix unique identification node and the heterogeneous matrix node to generate heterogeneous matrix node association data; A heterogeneous matrix architecture is established for the unique identification nodes of the heterogeneous matrix and the heterogeneous matrix nodes according to the heterogeneous matrix node association data to construct a heterogeneous matrix architecture.

7. The large-scale transport vehicle identification method based on radar and vision fusion according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: integrating the transport vehicle kinematic data and the transport vehicle transport characteristic data into dynamic heterogeneous transport vehicle data; Step S42: performing dynamic key frame analysis based on the dynamic heterogeneous data of the transport vehicle to generate key frame dynamic heterogeneous data of the transport vehicle; Step S43: performing time series analysis on the key frame transport vehicle dynamic heterogeneous data to generate time series key frame transport vehicle dynamic heterogeneous data, and performing nonlinear transport vehicle dynamic spatiotemporal feature analysis based on the time series key frame transport vehicle dynamic heterogeneous data to generate transport vehicle dynamic spatiotemporal feature data; Step S44: performing transport vehicle dynamic feature pattern recognition based on the transport vehicle dynamic spatiotemporal feature data to generate transport vehicle dynamic feature pattern data.

8. A large-scale transport vehicle identification system based on radar and vision fusion, characterized in that: Used to execute the large-scale transport vehicle identification method based on radar and vision fusion as claimed in claim 1, the large-scale transport vehicle identification system based on radar and vision fusion includes: The radar signal transport vehicle characteristic analysis module is used to obtain the data of the supervision area on the way of the large-scale transport vehicle; collect the transport vehicle monitoring radar signal on the way through the radar monitoring equipment and the data of the supervision area on the way of the large-scale transport vehicle to generate the transport vehicle monitoring radar signal on the way; analyze the transport vehicle characteristics of the radar signal according to the transport vehicle monitoring radar signal on the way to generate the transport vehicle characteristic data, wherein the transport vehicle characteristic data includes the transport vehicle static geometry data and the transport vehicle kinematic data; The transport vehicle visual feature analysis module is used to collect visual data of the transport vehicle on the way by using high-resolution monitoring equipment and data of the supervision area on the way of the large-scale transport vehicle, and generate visual data of the transport vehicle on the way; perform visual feature analysis on the transport vehicle on the way according to the visual data on the way of the transport vehicle, and generate visual feature data on the way of the transport vehicle, wherein the visual feature data on the way of the transport vehicle includes static transport vehicle appearance feature data and transport feature data of the transport vehicle; A transport vehicle external heterogeneous data mapping module is used to map the transport vehicle external heterogeneous data according to the transport vehicle static geometric data and static transport vehicle appearance feature data to generate a transport vehicle external heterogeneous data matrix; The transport vehicle dynamic feature pattern recognition module is used to perform transport vehicle dynamic feature pattern recognition based on the transport vehicle kinematic data and the transport feature data of the transport vehicle, and generate the transport vehicle dynamic feature pattern data; The transport vehicle identification feedback module is used to transmit the external heterogeneous data matrix of the transport vehicle and the corresponding dynamic characteristic pattern data of the transport vehicle to the terminal to perform the vehicle identification feedback operation of the large-scale transport vehicle.

Images