Traffic information early warning device and method

Abstract

The invention provides a traffic information early warning device and method, and belongs to the technical field of intelligent traffic, and the device comprises an equipment sensing module which is used for collecting radar data and video data related to traffic; the fusion module is used for extracting a video target from the video data, extracting a radar target from the radar data, and performing fusion tracking processing on the video target and the radar target to obtain target vector data; and the data module is in communication connection with the fusion module and is used for converting the target vector data into a message set in a target format and broadcasting the message set to the outside through a V2X road side communication unit arranged in the data module so as to carry out early warning. According to the invention, the radar data and the video data are fused and tracked to obtain the target vector data, the target vector data are converted into the message set in the target format, and the message set is broadcasted to the outside through the V2X module, so that the technical problem that the thunder-vision all-in-one machine cannot recognize traffic data and perform real-time early warning on traffic participants is solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a traffic information warning device and method. Background Art

[0002] With the rise of smart cities, smart transportation has become one of the key areas of urban development. The rapid development of informatization and technology has brought new opportunities to the field of smart transportation. The integrated laser vision machine is an advanced device that integrates advanced visual perception technology, big data analysis and intelligent decision-making. It can sense and analyze various data in the traffic environment, including vehicles, pedestrians, road conditions, etc., so as to provide accurate real-time information for the smart transportation system. The integrated laser vision machine cannot identify traffic data and provide real-time warnings to traffic participants. Summary of the invention

[0003] In view of this, it is necessary to provide a traffic information warning device and method to solve the technical problem that the radar-vision integrated machine cannot identify traffic data and provide real-time warnings to traffic participants.

[0004] In order to solve the above problems, on the one hand, the present invention provides a traffic information warning device, comprising: Device perception module, used to collect traffic-related radar data and video data; A fusion module, which is in communication with the device perception module, is used to extract video targets from the video data, extract radar targets from the radar data, and perform fusion tracking processing on the video targets and the radar targets to obtain target vector data; The data module is communicatively connected with the fusion module, and is used for converting the target vector data into a message set in a target format, and broadcasting the message set to the outside through its own built-in V2X roadside communication unit for early warning.

[0005] In a possible implementation, the data module includes: a structured data processing unit for converting target vector data into a message set in a target format; A V2X roadside communication unit, used to broadcast the message set to the on-board V2X device; The Ethernet communication unit is used to send the message set to the external device via Ethernet.

[0006] In a possible implementation, the V2X roadside communication unit includes: Satellite communications unit; Main antenna; Auxiliary antenna; The V2X communication module is used to broadcast the message set to the on-board V2X device through the satellite communication unit, the main antenna or the auxiliary antenna.

[0007] In a possible implementation, the device perception module includes: Radar, used to collect the spatial position and moving speed of traffic-related targets and generate radar data containing the spatial position and moving speed of the targets; The camera is used to collect raw traffic-related videos, identify targets in the raw videos, and generate video data containing target categories.

[0008] In a possible implementation, the camera includes: Lens; A CCD sensor, used for collecting raw traffic-related video through the lens; A video DSP processing chip is connected to the CCD sensor for preprocessing the original video to obtain a preprocessed video; An audio acquisition encoding unit, used to collect traffic-related audio data; The video encoding main control chip is respectively connected to the video DSP processing chip and the audio acquisition encoding unit for performing feature extraction and feature classification based on the pre-processed video and the audio data to realize target recognition and generate video data containing target categories.

[0009] In a possible implementation, the preprocessing includes: framing, color adjustment, white balance, contrast equalization, and image distortion.

[0010] In a possible implementation, the spatial position of the traffic-related target includes: the distance, position, and angle of the traffic-related target; The radar is used to perform 2D-FFT processing on each frame of data in each of its own RX channels, perform incoherent accumulation processing on the 2D-FFT processing results, and perform CFAR detection on the incoherent accumulation processing results to obtain the distance, position and moving speed of traffic-related targets, and also perform angle estimation based on data at corresponding RD spectrum positions on its own receiving channels through DBF or spatial spectrum estimation methods to obtain the angle of traffic-related targets.

[0011] It can be understood that DBF is a radar arrival angle estimation algorithm based on array signal processing. The DBF algorithm includes a DoA estimation and a beamforming algorithm.

[0012] In a possible implementation, the radar is a millimeter wave radar.

[0013] In a possible implementation, the fusion module includes: a computing kernel constructed by an ARM processor and an NPI processor; The computing kernel is used to extract video targets from video data and radar targets from radar data, and map the video targets to the radar coordinate system for fusion tracking and deep learning calculation with the radar targets to obtain target vector data.

[0014] On the other hand, the present invention further provides a traffic information warning method, which is applied to any of the above-mentioned warning devices, and the warning method comprises: Collect traffic-related radar data and video data based on the device perception module; Extract video targets from video data and radar targets from radar data based on the fusion module, and perform fusion tracking processing on the video targets and radar targets to obtain target vector data; Based on the data module, the target vector data is converted into a message set in the target format, and the message set is broadcast externally through its own built-in V2X roadside communication unit for early warning.

[0015] The beneficial effect of adopting the above-mentioned implementation method is: the traffic information warning device and method provided by the present invention, the warning device, includes: a device perception module, used to collect traffic-related radar data and video data; a fusion module, which is communicated with the device perception module, and is used to extract video targets from video data, and extract radar targets from radar data, and fuse the video targets with the radar targets for tracking and processing to obtain target vector data; a data module, which is communicated with the fusion module, and is used to convert the target vector data into a message set in a target format, and broadcast the message set to the outside through its own built-in V2X roadside communication unit for warning.

[0016] The present invention combines radar data with video data. Although radar data can measure speed, distance and angle, it cannot detect static targets and cannot accurately distinguish target types. Target detection using video data is easily affected by weather. By combining radar data with video data, these defects can be overcome, thereby accurately identifying traffic data and obtaining the number of target vectors. Finally, the data module converts the data into a message set in the target format, which is broadcast externally through the V2X module to achieve early warning. In summary, the present invention can solve the technical problem that the radar-visual integrated machine cannot identify traffic data and provide real-time early warnings to traffic participants. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 A functional block diagram of an embodiment of a traffic information warning device provided by the present invention; Figure 2 A system block diagram of another embodiment of the traffic information warning device provided by the present invention; Figure 3 A schematic diagram of another embodiment of the traffic information warning device provided by the present invention; Figure 4 A device data diagram of another embodiment of the traffic information warning device provided by the present invention; Figure 5 A flowchart of camera image data processing in an embodiment of the present invention; Figure 6 A structural diagram of a V2X module in an embodiment of the present invention; Figure 7 A data processing flow chart of a V2X module in an embodiment of the present invention; Figure 8 A diagram of equipment integration in another embodiment of the traffic information warning device provided by the present invention; Fig. 9 A data processing business logic diagram of another embodiment of the traffic information warning device provided by the present invention; Fig.10 The present invention provides a flow chart of an embodiment of the traffic information warning method. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present application, unless otherwise specified, “plurality” means two or more than two.

[0021] The terms "including" and "having" and any variations thereof in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product or equipment comprising a series of steps or modules is not necessarily limited to those steps or modules clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, methods, products or equipment.

[0022] The naming or numbering of the steps in the embodiments of the present invention does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The execution order of the named or numbered process steps can be changed according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0023] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0024] The present invention provides a traffic information warning device and method, which are respectively described below.

[0025] like Figure 1 As shown, the present invention provides a traffic information warning device, comprising: The device sensing module 101 is used to collect traffic-related radar data and video data; The fusion module 102 is connected to the device perception module 101 for extracting video targets from the video data and radar targets from the radar data, and fusing and tracking the video targets with the radar targets to obtain target vector data; The data module 103 is in communication connection with the fusion module 102, and is used for converting the target vector data into a message set in a target format, and broadcasting the message set to the outside through its own built-in V2X roadside communication unit for early warning.

[0026] It can be understood that the traffic information warning device based on V2X (Vehicle to Everything, communication between the vehicle-mounted unit and other devices) proposed in the present invention is a V2X integrated traffic information warning device based on radar vision (abbreviated as: V2X integrated machine), which provides more reliable perception data support for the V2X road test communication unit.

[0027] In order to achieve the above purpose, the technical solution adopted by the present invention is: The integrated radar and vision device is integrated with the V2X road side communication unit RSU (Road Side Unit) device to form a new integrated device, which includes the road side road condition detection function of the integrated radar and vision device and the message broadcast warning function of the V2X road side communication unit RSU.

[0028] Compared with the existing method, the device provided by the present invention has the following beneficial effects: V2X is widely used in vehicle-road collaboration projects and intelligent transportation fields to provide reliable application needs for solving smart travel problems. It combines V2X technology with millimeter-wave radar and video in the transportation field. Millimeter-wave radar can measure speed, distance, and angle, but it cannot detect static targets and the results cannot be presented in an intuitive image, and it cannot accurately distinguish the target type; video is easily affected by weather, and the integrated radar and vision machine (radar and vision sensing device) can just solve these problems. However, traditional radar and vision sensing devices cannot process and identify traffic data and provide real-time warning information to traffic participants.

[0029] In the existing V2X solutions, roadside equipment basically needs to be installed with cameras, radars, and V2X roadside communication units RSU to realize the basic application functions of V2X. The present invention only needs to install one device to realize post-road testing, which can greatly reduce equipment costs, construction and installation costs, and later maintenance costs.

[0030] In the existing V2X solution, multiple sensing devices are combined with the V2X roadside communication unit RSU to perform data sensing, processing and message sending. The present invention can perform internal data docking and timely message broadcasting for the sensing data, thereby improving the efficiency and timeliness of data utilization and reducing the impact caused by the lack of data from other sensing devices.

[0031] Based on the advanced visual perception, big data analysis and intelligent decision-making technologies of the integrated radar and vision device, it can obtain the precise vehicle trajectory data in real time, the vehicle speed, heading angle and movement trend at the intersection, and send the road-side data to the vehicle-side equipment in time through the integrated V2X roadside communication unit RSU. The present invention has the following beneficial effects: Highly integrated all-in-one equipment can be used in multiple scenarios. The integrated equipment ensures the efficiency and stability of data use, reduces equipment procurement costs, construction costs and later equipment maintenance costs, and can greatly increase the speed of vehicle networking project construction.

[0032] Real-time data has low latency and efficient transmission, and can improve the consistency between the collected real-time data of road tests and the data sent down.

[0033] In some embodiments, the data module includes: a structured data processing unit for converting target vector data into a message set in a target format; V2X roadside communication unit, used to broadcast message sets to vehicle-mounted V2X devices.

[0034] The Ethernet communication unit is used to send the message set to the external device via Ethernet.

[0035] In some embodiments, the V2X roadside communication unit includes: Satellite communications unit; Main antenna; Auxiliary antenna; The V2X communication module is used to broadcast the message set to the on-board V2X device through the satellite communication unit, the main antenna or the auxiliary antenna.

[0036] In some embodiments, the device perception module includes: Radar, used to collect the spatial position and moving speed of traffic-related targets and generate radar data containing the spatial position and moving speed of the targets; The camera is used to collect raw traffic-related videos, identify targets in the raw videos, and generate video data containing target categories.

[0037] In some embodiments, the camera comprises: Lens; a CCD (charge coupled device) sensor for collecting raw traffic-related video through the lens; A video DSP processing chip is connected to the CCD sensor for preprocessing the original video to obtain a preprocessed video; An audio acquisition encoding unit, used to collect traffic-related audio data; The video encoding main control chip is respectively connected to the video DSP processing chip and the audio acquisition encoding unit for performing feature extraction and feature classification based on the pre-processed video and the audio data to realize target recognition and generate video data containing target categories.

[0038] The preprocessing includes: framing, color adjustment, white balance, contrast balance and image distortion.

[0039] In some embodiments, the spatial position of the traffic-related target includes: the distance, position, and angle of the traffic-related target; The radar is used to perform 2D-FFT processing on each frame of data in each of its own RX channels, and perform incoherent accumulation processing on the 2D-FFT processing results, and perform CFAR detection on the incoherent accumulation processing results to obtain the distance, position and moving speed of traffic-related targets, and also perform angle estimation based on the data at the corresponding RD spectrum position on its own receiving channel through DBF or spatial spectrum estimation method to obtain the angle of the traffic-related target. The radar is a millimeter wave radar.

[0040] In some embodiments, the fusion module includes: a computing kernel constructed by an ARM processor and an NPI processor; The computing kernel is used to extract video targets from video data and radar targets from radar data, and map the video targets to the radar coordinate system for fusion tracking and deep learning calculation with the radar targets to obtain target vector data.

[0041] In some embodiments, the V2X integrated device provided by the present invention is a roadside perception unit for smart transportation. The radar module adopts embedded tools and interface debugging methods, supports bidirectional 12 lanes, and has a maximum detection distance of 400 meters and a message broadcast distance of 500 meters. The fusion detection accuracy is ≥98%, and the distance resolution is ≤0.25m. The device uses a color 8 million 4K camera and can work under complex and harsh weather conditions. The device has a built-in low-power edge computing core with ARM+NPU architecture, which can classify and identify traffic target types and behavior events more accurately through deep learning. The device supports target classification and recognition, supports V2X message set broadcasting, supports scene customization, etc.

[0042] The V2X all-in-one device provided by the present invention has highly reliable V2X functions, wide-area microwave detection technology, low-illumination wide-angle camera, radar video deep fusion technology, built-in artificial intelligence self-calibration technology, integrated industrial-grade design, and highly integrated structure.

[0043] The equipment parameters of the V2X integrated device provided by the present invention are shown in Table 1: Table 1: Equipment parameters of V2X integrated device

[0044] The system block diagram of the V2X integrated device in this embodiment is as follows Figure 2 As shown in the figure, the schematic diagram of the V2X integrated device is as follows Figure 3 As shown, the corresponding device data diagram is as follows Figure 4 shown.

[0045] The V2X all-in-one machine is an intelligent sensor terminal device designed specifically for smart traffic vehicle-road collaboration. It is a traffic sensor with high-performance computing that integrates a camera, radar, and V2X. The radar-vision fusion simultaneously connects the original video stream and radar data stream to the embedded processor in the all-in-one machine through the protocol interface. The built-in processor directly extracts the target from the original video stream, and then projects the video target into the radar coordinate system through the built-in coordinate mapping system. Finally, the video target and radar target are fused and tracked to achieve real-time vectorization and tracking of global targets, and convert the traffic information after fusion processing and decision-making into structured data.

[0046] The V2X all-in-one converts the fused video and radar data in real time through the built-in ARM+NPU (network processor) computing core, outputs target vector data, and forms a global real-time target data source for intersections or sections by performing deep learning calculations on multi-directional vector target data, providing data sources for MEC processing, decision-making, and control through standard protocol interfaces. It can also directly broadcast externally in the form of message sets through the V2X module of the all-in-one itself, communicate and interact with the on-board V2X equipment, and send the information of traffic participants at the entire traffic intersection to the surrounding intelligent networked vehicle OBUs (on-board units).

[0047] The V2X all-in-one device has a built-in GNSS antenna and can be equipped with a built-in intersection or road section map. It can be connected to external signal machines and MEC devices through Ethernet, and supports access and communication of most multi-source perception and signal devices on the market according to standard interface protocols. Its own video module can detect the types of traffic participants, such as vehicles, pedestrians and traffic participants. The radar module can accurately detect the location, speed and other information of the target. The V2X module can timely send the traffic information identified and processed by the all-in-one device to the external warning.

[0048] The V2X integrated machine product provided by the present invention mainly includes the following modules: system power module, SOC system module, video module, V2X module, radar module, WIFI / BT (Bluetooth) module, HSM module (hardware security module), 1000M Ethernet module. In addition, the V2X integrated machine also includes other modules, such as: indicator light circuit, debug serial port circuit.

[0049] The radar module provided by the present invention has a radar imaging principle that uses the scattering and reflection characteristics of electromagnetic waves to obtain the spatial position and morphological information of the target. The radar system transmits a pulse signal, which is scattered and reflected by the target. The received signal is analyzed and processed to finally form an image of the target. By measuring parameters such as the signal delay, amplitude and phase, the distance, direction, height and moving speed of the target are obtained. This information can be displayed and analyzed in real-time or offline mode for applications such as target identification, tracking and positioning.

[0050] The specific steps of radar signal processing include: 1) 2D-FFT: First, perform 2D-FFT processing (RD processing) on ​​the data in each Frame of each RX channel to obtain the range-Doppler spectrum of each channel. This 2D-FFT spectrum is a complex result and cannot directly extract the target position and speed information. Therefore, incoherent accumulation processing is required.

[0051] 2) Incoherent accumulation: that is, take the modulus square value of the 2D-FFT result corresponding to each channel, then sum and average the corresponding positions of the 2D-FFT modulus square results of all channels to obtain a synthesized RD spectrum. This completes the incoherent accumulation process. Each peak on the synthesized RD spectrum represents a possible target or interference. There are two goals of incoherent accumulation: first, to improve the signal-to-noise ratio and improve detection accuracy; second, to smooth the noise and reduce the false detection rate in the CFAR (constant false alarm rate) stage.

[0052] 3) CFAR detection: Perform constant false alarm detection based on the incoherent accumulation results to obtain all possible targets in this frame. All detected targets are not necessarily real targets, and may be strong clutter interference. The detected interference targets can be filtered out during the target tracking phase. According to the corresponding position of the possible targets detected by CFAR on the RD spectrum, the distance and speed information of the target can be calculated.

[0053] 4) Angle estimation: After CFAR is completed, although the position and speed information of each possible target are obtained, the angle information of the target is not yet obtained. In order to obtain the angle information of each target, it is necessary to use the data of the corresponding RD spectrum position of each target on all receiving channels, and perform angle estimation through DBF or spatial spectrum estimation method to obtain the target angle information.

[0054] The above steps complete the calculation of the distance, speed and angle information of each target. This information can be used for clustering and tracking, and further obtain the target's trajectory information as well as size and type information.

[0055] The camera module provided by the present invention can refer to Figure 5 As shown in the figure, its workflow is: image input - preprocessing - feature extraction - feature classification - matching - complete recognition, that is, input the data of the camera, perform detection, classification, segmentation and other calculations based on each frame of information, and finally use multiple frames of information to track the target and output relevant results; including: 1) Preprocessing includes framing, color adjustment, white balance, contrast balance, image distortion, etc.; 2) Feature extraction extracts feature points from the image based on preprocessing; 3) Target recognition is based on the output of feature data to identify and classify objects in the image - people, cars, traffic signs, etc., using algorithms such as machine learning and neural networks.

[0056] The V2X module includes a computing unit, a storage unit, an encryption unit, a main communication unit, an antenna unit, a GNSS (Global Navigation Satellite System) unit, and a power supply unit. It supports 4G / 5GUU and C-V2X PC5 air interfaces concurrently, and realizes direct communication between vehicles and roads through the PC5 interface. The main communication unit can be Figure 6 For V2X communication as shown, the antenna unit may include Figure 6 The main antenna and auxiliary antenna shown, GNSS can include Figure 6 The data processing flow chart of the Beidou navigation system or GPS navigation system shown in the figure is as follows: Figure 7 shown.

[0057] C-V2X (Cellular Vehicle-to-Everything) is a cellular network-based vehicle wireless communication technology that enables all-round communication between vehicles and their surroundings (including other vehicles, infrastructure, pedestrians, and networks).

[0058] PC5 refers to the direct communication interface, a short-distance direct communication interface between vehicles, people and road infrastructure. Its characteristics are: low-latency, high-capacity and high-reliability communication through direct connection, broadcasting and network scheduling.

[0059] UU refers to the cellular network communication interface, the communication interface between the terminal and the base station, and its characteristics are: achieving reliable communication over long distances and a larger range.

[0060] Finally, the above modules are integrated into the V2X integrated device as shown in the following diagram: Figure 8 As shown, the corresponding data processing business logic diagram is as follows Fig. 9 shown.

[0061] The present invention also provides a traffic information warning method, which is applied to the above-mentioned warning device. Fig.10 As shown, the early warning method includes: S1001. Collecting traffic-related radar data and video data based on a device perception module; S1002, extracting a video target from the video data and a radar target from the radar data based on a fusion module, and performing fusion tracking processing on the video target and the radar target to obtain target vector data; S1003. Convert the target vector data into a message set in a target format based on the data module, and broadcast the message set to the outside through the built-in V2X roadside communication unit for early warning.

[0062] The traffic information warning device and method provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A traffic information warning device, characterized in that: include: Device perception module, used to collect traffic-related radar data and video data; A fusion module, which is in communication with the device perception module, is used to extract video targets from the video data, extract radar targets from the radar data, and perform fusion tracking processing on the video targets and the radar targets to obtain target vector data; The data module is communicatively connected with the fusion module, and is used for converting the target vector data into a message set in a target format, and broadcasting the message set to the outside through its own built-in V2X roadside communication unit for early warning.

2. The traffic information warning device according to claim 1, characterized in that: The data module comprises: a structured data processing unit for converting target vector data into a message set in a target format; A V2X roadside communication unit, used to broadcast the message set to the on-board V2X device; The Ethernet communication unit is used to send the message set to the external device via Ethernet.

3. The traffic information warning device according to claim 2, characterized in that: The V2X roadside communication unit comprises: Satellite communications unit; Main antenna; Auxiliary antenna; The V2X communication module is used to broadcast the message set to the on-board V2X device through the satellite communication unit, the main antenna or the auxiliary antenna.

4. The traffic information warning device according to claim 1, characterized in that: The device perception module includes: Radar, used to collect the spatial position and moving speed of traffic-related targets and generate radar data containing the spatial position and moving speed of the targets; The camera is used to collect raw traffic-related videos, identify targets in the raw videos, and generate video data containing target categories.

5. The traffic information warning device according to claim 4, characterized in that: The camera comprises: Lens; A CCD sensor, used for collecting raw traffic-related video through the lens; A video DSP processing chip is connected to the CCD sensor for preprocessing the original video to obtain a preprocessed video; An audio acquisition encoding unit, used to collect traffic-related audio data; The video encoding main control chip is respectively connected to the video DSP processing chip and the audio acquisition encoding unit for performing feature extraction and feature classification based on the pre-processed video and the audio data to realize target recognition and generate video data containing target categories.

6. The traffic information warning device according to claim 5, characterized in that: The preprocessing includes: framing, color adjustment, white balance, contrast equalization and image distortion.

7. The traffic information warning device according to claim 4, characterized in that: The spatial location of traffic-related targets, including: the distance, position and angle of traffic-related targets; The radar is used to perform 2D-FFT processing on each frame of data in each of its own RX channels, perform incoherent accumulation processing on the 2D-FFT processing results, and perform CFAR detection on the incoherent accumulation processing results to obtain the distance, position and moving speed of traffic-related targets, and also perform angle estimation based on data at corresponding RD spectrum positions on its own receiving channels through DBF or spatial spectrum estimation methods to obtain the angle of traffic-related targets.

8. The traffic information warning device according to claim 4, characterized in that: The radar is a millimeter wave radar.

9. The traffic information warning device according to any one of claims 1 to 8, characterized in that: The fusion module includes: a computing kernel constructed by an ARM processor and an NPI processor; The computing kernel is used to extract video targets from video data and radar targets from radar data, and map the video targets to the radar coordinate system for fusion tracking and deep learning calculation with the radar targets to obtain target vector data.

10. A traffic information warning method, characterized in that: The early warning method is applied to the early warning device according to any one of claims 1 to 9, and the early warning method includes: Collect traffic-related radar data and video data based on the device perception module; Extract video targets from video data and radar targets from radar data based on the fusion module, and perform fusion tracking processing on the video targets and radar targets to obtain target vector data; Based on the data module, the target vector data is converted into a message set in the target format, and the message set is broadcast externally through its own built-in V2X roadside communication unit for early warning.

Images