Traffic warning system, computer-readable medium using the same and operation method thereof

Abstract

The traffic warning system includes a visual sensor, a radar sensor, and a processor. The visual sensor detects visual data for each of at least one object. The radar sensor detects radar data for each of at least one object. The processor analyzes the radar sensor data to obtain radar confidence level; analyzes the visual sensor data to obtain visual confidence level; and if both radar confidence level and visual confidence level are greater than preset values, analyzes the visual sensor data and radar sensor data to determine the position and speed of at least one object.

Description

[Technical Field]

[0001] This disclosure relates to a traffic warning system, a computer-readable medium using the system, and a method of operation thereof. [Previous Technology]

[0002] With the increase in the number and types of vehicles, the probability of traffic accidents is also increasing year by year. Clearly, in addition to the continuous advancements in power management systems in transportation technology, improving safety during driving is another crucial issue. For example, there are many objects on a bicycle's route that riders need to pay attention to, such as traffic signs, pedestrians, and obstacles. Riders also need to be aware of the road environment as a basis for their next driving actions.

[0003] Nowadays, traffic accidents often occur because riders do not pay attention to road conditions. Riders may fail to notice traffic signs or obstacles on the road due to fatigue or distraction, and thus be unable to make correct driving actions based on the traffic signs or obstacles ahead. [Summary of the Invention]

[0004] This disclosure provides an embodiment of a traffic warning system. The traffic warning system includes a vision sensor, a radar sensor, and a processor. The vision sensor is used to detect visual sensing data of at least one object. The radar sensor is used to detect radar sensing data of at least one object. The processor is electrically connected to the vision sensor and the radar sensor and is used to: obtain a radar confidence level of the radar sensing data by analyzing the radar sensing data; obtain a visual confidence level of the visual sensing data by analyzing the visual sensing data; and if both the radar confidence level and the visual confidence level are greater than a preset value, obtain the position and speed of at least one object by analyzing the visual sensing data and the radar sensing data.

[0005] In one example, the preset value is between 70% and 90%.

[0006] In one example, the processor is used to: determine the number of at least one object approaching the traffic warning system if both radar confidence and visual confidence are greater than preset values.

[0007] In one example, the processor is used to: output a warning signal if the number of at least one object is not greater than a preset number.

[0008] In one example, the processor is used to: if the number of at least one object is greater than a preset number, output a warning signal for the object closest to the traffic warning system.

[0009] In one example, the processor is used to: if the radar confidence level and the visual confidence level are not greater than a preset value and the radar confidence level is higher than the visual confidence level, determine the number of at least one object approaching the traffic warning system based on the radar sensing data.

[0010] In one example, the processor is used to: if the radar confidence and visual confidence are not greater than preset values ​​and the visual confidence is higher than the radar confidence, determine the number of at least one object approaching the traffic warning system based on the visual sensing data and the radar sensing data.

[0011] In one example, the processor is used to: output a warning signal if the number of at least one object is not greater than a preset number.

[0012] In one example, the processor is used to: if the number of at least one object is greater than a preset number, output a warning signal for the object closest to the traffic warning system.

[0013] In one example, the processor is used to: determine whether at least one object is detected; and if no object is detected, keep the radar sensor on and keep the vision sensor off.

[0014] In one example, the processor is used to: determine whether at least one object is detected; and, if the radar sensor detects at least one object, activate the visual sensor.

[0015] In one example, the processor is used to periodically turn on the vision sensor.

[0016] Another embodiment of this disclosure provides an operating method for a traffic warning system. The operating method includes the following steps: a vision sensor detects visual sensing data of at least one object; a radar sensor detects radar sensing data of at least one object; a processor obtains radar confidence level of the radar sensing data by analyzing the radar sensing data; the processor obtains visual confidence level of the visual sensing data by analyzing the visual sensing data; and if both the radar confidence level and the visual confidence level are greater than a preset value, the processor obtains the position and speed of at least one object by analyzing the visual sensing data and the radar sensing data.

[0017] In one example, the preset value is between 70% and 90%.

[0018] In one example, the operation method further includes: if both radar confidence and visual confidence are greater than preset values, determining the number of at least one object approaching the traffic warning system.

[0019] In one example, the operation method further includes: if the number of at least one object is not greater than a preset number, the processor outputs a warning signal.

[0020] In one example, the operation method further includes: if the number of at least one object is greater than a preset number, the processor outputs a warning signal for the object closest to the traffic warning system.

[0021] In one example, the operation method further includes: if the radar confidence and visual confidence are not greater than a preset value and the visual confidence is higher than the radar confidence, the processor determines the number of at least one object approaching the traffic warning system based on the visual sensing data and the radar sensing data.

[0022] In one example, the operation method further includes: if the number of at least one object is not greater than a preset number, the processor outputs a warning signal.

[0023] In one example, the operation method further includes: if the number of at least one object is greater than the preset number, the processor outputs a warning signal for the object closest to the traffic warning system.

[0024] In one example, the operation method further includes: the processor determining whether at least one object is detected; and if no object is detected, the processor keeps the radar sensor on and keeps the vision sensor off.

[0025] In one example, the operation method further includes: the processor determining whether at least one object is detected; and if the radar sensor detects at least one object, the processor activating the visual sensor.

[0026] In one example, the operation method further includes: periodically turning on the visual sensor.

[0027] Another embodiment of this disclosure provides a computer-readable medium. The computer-readable medium is a non-transitory computer-readable medium storing a plurality of instructions executed by at least one processor of a traffic warning system to cause the traffic warning system to perform the following methods: detecting visual sensing data of at least one object; detecting radar sensing data of at least one object; obtaining a radar confidence level of the radar sensing data by analyzing the radar sensing data; obtaining a visual confidence level of the visual sensing data by analyzing the visual sensing data; and, if both the radar confidence level and the visual confidence level are greater than a preset value, obtaining the position and speed of at least one object by analyzing the visual sensing data and the radar sensing data.

[0028] In one example, a non-transitory computer-readable medium stores these instructions, which are executed by at least one processor of the traffic warning system to cause the traffic warning system to perform another method as described in any of the preceding examples.

[0029] In order to better understand the above and other aspects of this disclosure, specific embodiments are described below in detail with reference to the accompanying drawings:

Implementation Method

[0031] This disclosure proposes a traffic warning system, a computer-readable medium using the system, and a method for operating the system. It integrates two types of sensors, such as radar or microwave sensors and visual sensors. Through feasible operating methods and control logic, it fully utilizes the advantages of both, significantly reducing the power consumption required for system operation. In particular, it can be applied to various mobile light vehicles (such as bicycles, electric bicycles, or other multi-wheeled personal mobility solutions) to provide real-time warnings, assisting riders in improving riding safety and avoiding potential dangers from approaching vehicles or threats from the rider's blind spots. The detection results of such emergency situations can be detected by the system, and the initial information is promptly provided to the rider, thereby achieving beneficial effects such as power saving, low interference, and accurate warnings.

[0032] Please refer to Figure 1, which illustrates a functional block diagram of a traffic warning system 100 according to an embodiment of the present invention. The traffic warning system 100 can be configured in a vehicle (e.g., a bicycle, an electric bicycle, or other multi-wheeled personal mobility solution).

[0033] As shown in Figure 1, the traffic warning system 100 includes a vision sensor 110, a radar sensor 120, and a processor 130. The vision sensor 110 is used to detect visual sensing data DV for each of at least one object (or object) 10. The radar sensor 120 is used to detect radar sensing data DR for each of at least one object 10. The processor 130 is used to obtain the radar confidence level LR of the radar sensing data by analyzing the radar sensing data DR; and to obtain the visual confidence level LV of the visual sensing data DV by analyzing the visual sensing data DV; if both the radar confidence level LR and the visual confidence level LV are greater than preset values, the position and speed of the object 10 are obtained by analyzing the visual sensing data DV and the radar sensing data DR. In this embodiment, the vision sensor 110 and the radar sensor 120 can work together to improve the accuracy of object detection.

[0034] The "visual sensor" in this article can be implemented through an image capturing device, such as a camera (e.g., a GoPro camera mounted on a helmet, a wide-angle camera, an infrared camera, etc.). In addition, the traffic warning system 100 (e.g., a visual sensor 110 or a processor 130) can identify the object 10 by using an artificial intelligence (AI) model. The AI ​​model can be obtained in advance through various suitable machine learning techniques, such as deep learning. The object 10 is, for example, an obstacle, a vehicle, a pedestrian, etc. The visual sensor 110 has the following characteristics: (1) a close and near sensing range, which can reduce blind spots in the field of view (FOV) area, wherein the field of view area has a length of 30 meters and a field of view of 144 degrees; and (2) the visual sensor can detect multiple targets simultaneously (e.g., more than 10 target objects).

[0035] The radar sensor 120 is, for example, a 24G radar, which has the lowest power consumption and is selectable from different specifications with or without lane recognition capability. Generally, the power consumption of a visual sensor, for example 10W, is much higher than that of a radar sensor, which leads to limitations in applicability. In this embodiment, the traffic warning system can integrate radar and visual sensors, minimizing power consumption and further improving so-called "smart" functionality (i.e., lane recognition capability with adaptive energy-saving schemes or mechanisms). Radar (or microwave) sensors have the characteristics of long and wide sensing distances, and by utilizing, for example, 24 GHz millimeter-wave, the coverage can be further extended beyond a visual area at least 100 meters (m) in length and at least 8 meters in width.

[0036] The processor 130 can be electrically connected to the vision sensor 110 and the radar sensor 120, and is used to control the sensors to turn on and off according to different traffic scenarios or environmental conditions, and to process the data sensed or generated by the sensors. The processor 130 can control the vision sensor 110 and the radar sensor 120 to perform the operation method of the traffic warning system 100. The processor 130 can obtain the position and size of the object 10 (or object frame) based on the visual sensing data DV, and obtain the position and speed of the object 10 based on the radar sensing data DR.

[0037] Please refer to Figures 2A and 2B. Figures 2A and 2B illustrate the operation method of the traffic warning system 100 in Figure 1, corresponding to different situations in a continuous flowchart.

[0038] In step S110, the radar sensor 120 remains on. Furthermore, if no object is detected during the detection cycle, the radar sensor 120 may remain on, while the vision sensor 110 may remain off to save power.

[0039] In step S120, the radar sensor 120 can detect whether any object 10 is present around the traffic warning system 100. If any object 10 is present around the traffic warning system 100, the process proceeds to step S140A.

[0040] In step S130A, the radar sensor 120 remains on, while the vision sensor 110 remains off to save power.

[0041] In step S130B, at each sampling interval (e.g., 2 to 5 seconds), the visual sensor 110 is activated to detect whether any object 10 is present around the traffic warning system 100. In one embodiment, the visual sensor 110 may be activated for an enable time (e.g., 2 to 5 seconds) at each sampling interval. In one embodiment, the enable time may be less than, greater than, or equal to the sampling interval.

[0042] In step S140A, if any object is detected, the vision sensor 110 is turned on to increase or improve the confidence of the detection result. The vision sensor 110 can detect visual sensing data DV of at least one object 10.

[0043] In steps S140B and S140C, the processor 130 executes a fusion procedure. In the fusion procedure, the processor 130 can obtain the radar confidence level (LR) of the radar sensing data DR by analyzing the radar sensing data DR (e.g., based on reflected signal strength, time of flight, etc.), and obtain the visual confidence level (LV) of the visual sensing data DV by analyzing the visual sensing data DV (e.g., based on captured images and image processing results, etc.). In one embodiment, the radar confidence level LR can be obtained using, for example, millimeter-wave radar, while the visual confidence level LV can be obtained using, for example, a CMOS (Complementary Metal-Oxide-Semiconductor) sensing camera, a QVGA (Quarter Video Graphics Array) sensing camera, and / or an infrared sensing camera.

[0044] In one embodiment, only one of the radar sensor and the visual sensor may be operational, providing more reliable detection information under certain special circumstances. In the first case, the accuracy of visual sensing drops sharply in adverse weather conditions and / or nighttime environments, so only the radar sensor can detect vehicles and further improve the situation under adverse conditions. In the second case, when an oncoming object is beyond the detection range of the radar sensor, the visual sensor can detect the vehicle located within the radar blind spot, further improving the situation of loss. Therefore, the two sensors can operate adaptively according to different situations.

[0045] In step S140D, the processor 130 determines whether both the radar confidence level (LR) and the visual confidence level (LV) are greater than preset values. If both the radar confidence level (LR) and the visual confidence level (LV) are greater than preset values, the process proceeds to step S150A. If both the radar confidence level (LR) and the visual confidence level (LV) are not greater than preset values, the process proceeds to step S160. In one embodiment, the preset value may be between, for example, 70% and 90% (e.g., 80%), or even smaller or larger. The preset value may be an indicative value or a percentage to reflect the quality or reliability of the data.

[0046] In step S150A, the processor 130 can obtain the position (e.g., relative position) and velocity (e.g., relative velocity) of the object 10 by analyzing visual sensing data DV and radar sensing data DR.

[0047] In step S150B, the processor 130 may determine the number of at least one object 10 approaching (e.g., the relative position and relative speed of the object 10 are positive) the traffic warning system 100.

[0048] In step S160A, the processor 130 can determine whether the radar confidence level LR is higher than the visual confidence level LV. If the radar confidence level LR is higher than the visual confidence level LV, the process proceeds to step S160B. If the radar confidence level LR is not higher than the visual confidence level LV (for example, the visual confidence level LV is higher than the radar confidence level LR), the process proceeds to step S160C.

[0049] In step S160B, the processor 130 determines the number of objects 10 approaching the traffic warning system based on radar sensing data DR (e.g., position and speed).

[0050] In step S160C, the processor 130 determines the number of objects 10 approaching the traffic warning system based on visual sensing data DV (e.g., position) and radar sensing data DR (e.g., speed or speed estimation).

[0051] In step S170, the processor 130 determines whether the number of objects 10 is not greater than a preset number. If the number of objects 10 is not greater than the preset number, the process proceeds to step S180A. If the number of objects 10 is greater than the preset number, the process proceeds to step S180B. In one embodiment, the preset number is, for example, 5, or even fewer or more.

[0052] In step S180A, the processor 130 outputs an alarm signal S1. The alarm signal S1 may be, for example, sound, vibration, image, light, etc.

[0053] In step S180B, the processor 130 outputs a warning signal S1 to the object 10 closest to the traffic warning system 100. For example, when the distance between the object 10 closest to the traffic warning system 100 and the traffic warning system 100 is equal to or less than a preset distance, the processor 130 outputs a warning signal S1.

[0054] Please refer to Figure 3, which illustrates the present disclosure of object recognition for different lanes based on visual algorithms.

[0055] First, the processor 130 obtains object frame 11 (or vehicle recognition frame) information based on the size of the approaching object. Then, the processor 130 calculates the size pixels based on different types of objects. Then, the processor 130 calculates the object position relative to the rider based on the calculated size pixels (e.g., the object position can be represented by distance distH and / or distance distW), and the value of the object position of object 10 is a reference value for different types of vehicles. DistW is the distance between object 10 and the edge of the lane (or roadside), and distance distH is the distance between the traffic warning system 100 and object 10. Then, the value of distW is used to determine whether the detected vehicle is in the same lane as the rider. For example, if distW is less than 2 meters, the processor 130 of the traffic warning system 100 determines that the approaching object is located in the same lane as the rider; if distW is not less than 2 meters, the processor 130 determines that the approaching object is located in a different lane.

[0056] Please refer to Figure 4, which is a schematic diagram showing the relative relationship between the relative speed (horizontal axis) of object 10 and traffic warning system 100 and the stopping distance (vertical axis) corresponding to a specific relative speed. The stopping distance is: the distance required or estimated for an approaching vehicle to come to a complete stop from a moving state when it approaches the vehicle of a rider equipped with traffic warning system 100 at different relative speeds.

[0057] As shown in Figure 4, compared to the warning mode, the stopping distance in hazard mode is shorter at the same relative speed. Compared to normal mode, the stopping distance in warning mode is shorter at the same relative speed. Based on the above observations, this means that hazard mode is more urgent than warning mode, therefore, if any situation is sensed to reach or exceed the level of hazard mode, a warning should be issued to the rider with the highest priority.

[0058] When an approaching object comes from behind the rider, the object 10 will be identified as being in the collision zone, and depending on the remaining reaction time, the rider will be notified with a warning message or a danger message. Furthermore, when a potentially dangerous object is in the threat zone, the rider will be notified in a warning mode. The collision zone and threat zone are used here for illustrative or exemplary purposes only.

[0059] As shown in Figures 2A, 2B, and 4, when the number of vehicles is not greater than a preset number (e.g., 5), the process proceeds to step 180A. When the number of vehicles is greater than the preset number, the process proceeds to step 180B. In step 180A, the processor can output warning signals according to the various modes (i.e., normal mode, warning mode, and / or danger mode). In step 180B, the processor can detect the first "preset number" dangerous vehicles (e.g., the first 5 dangerous vehicles) and output warning signals S1 in different modes accordingly. Dangerous vehicles refer to vehicles that are very close to (or continuously approaching) the traffic warning system 100 and may collide with it in a very short time.

[0060] Please refer to Figures 5 and 6. Figure 5 shows a schematic diagram of the collision area CZ and the threat area TZ in the embodiment, while Figure 6 shows a schematic diagram of the relative speed (horizontal axis) of the object 10 relative to the traffic warning system 100 and the corresponding parking distance (vertical axis) according to another embodiment.

[0061] As shown in Figure 5, the area where the traffic warning system 100 is located is defined as the collision zone CZ. A reference line R1 passes through the traffic warning system 100, and the collision zone CZ has a first collision boundary CZ1 and a second collision boundary CZ2. The first collision width WCZ1 is defined between the reference line R1 and the first collision boundary CZ1, and the second collision width WCZ2 is defined between the reference line R1 and the second collision boundary CZ2. In an embodiment, the range of the first collision width WCZ1 may be, for example, between 1 meter and 3 meters (e.g., 2 meters), and the range of the second collision width WCZ2 may be, for example, between 1 meter and 3 meters (e.g., 2 meters). The first collision width WCZ1 and the second collision width WCZ2 may be the same or different. When an approaching object is identified as being located within the collision zone CZ, the processor 130 can use the relative relationship shown in Figure 4 to determine the pattern.

[0062] As shown in Figure 5, the threat zone TZ is located outside the collision zone CZ. Furthermore, multiple threat zones TZ are connected to and adjacent to the first collision boundary CZ1 and the second collision boundary CZ2, respectively. The threat zone TZ has a threat width WTZ, which can range from, for example, 1 meter to 3 meters (e.g., 2 meters). When an approaching object is identified as being located within the threat zone TZ (outside the collision zone CZ), the processor 130 can determine the pattern using the relative relationship shown in Figure 6. Furthermore, the situation is more urgent when the approaching vehicle is located within the collision zone CZ. The traffic warning system 100 can operate appropriately according to different traffic conditions and different patterns, and adaptively output warning signals accordingly.

[0063] As shown in Figure 5, compared to the radar sensing range RR of radar sensor 120, the visual sensing range RV of vision sensor 110 has a wider sensing angle, which can reduce and eliminate blind spots near traffic warning system 100. Compared to the visual sensing range RV of vision sensor 110, the radar sensing range RR of radar sensor 120 has a longer sensing length, which can increase the detectable area far from traffic warning system 100. Furthermore, in one embodiment, the radar sensing length DS R of radar sensor 120 is greater than the visual sensing length DS V of vision sensor 110. In the embodiment, the radar sensing length DS R can be between, for example, 90 meters and 110 meters (e.g., 100 meters), and the visual sensing length DS V can be between, for example, 20 meters and 40 meters (e.g., 30 meters).

[0064] As shown in Table 1 below, it presents schematic diagrams of various modes of different embodiments of this disclosure. In mode 1, the radar sensor is always (or remains) on, while the vision sensor is periodically on (e.g., step S130B in Figure 2A), thus reducing power consumption. In mode 2, when the radar sensor detects a high-speed approaching object, the vision sensor is immediately turned on due to a potential hazard (e.g., step S140A in Figure 2A). In mode 3, the radar sensor is always on during severe weather (e.g., heavy rain, heavy fog), while the vision sensor is off, or the vision sensor is off when the visual confidence level (LV) frequently falls below a value between 50% and 80%. In mode 4, when an object behind the rider stops for more than a few seconds (e.g., 2 to 5 seconds), the radar sensor is temporarily or periodically off, and the vision sensor is off.

[0065] Table 1 radar sensor visual sensors Mode 1 Keep it on Turn on periodically (e.g., every 2 seconds, 5 seconds, etc.) Mode 2 High-speed object detected approaching Activate immediately upon the occurrence of a potential risk event. Mode 3 Keep it on Turn off the power in severe weather (such as heavy rain or fog) or when visual credibility frequently falls below 50%. Mode 4 When the object behind the rider stops for more than a few seconds (e.g., 2 to 5 seconds), a temporary or periodic power outage occurs. Turn off the power.

[0066] The traffic warning system disclosed herein, wherein a computer-readable medium storing executable instructions is executed by at least one processor of the traffic warning system using relevant algorithms or methods, includes at least the following features:

[0067] 1. In order to utilize radar and vision sensing technologies simultaneously, one of them is adaptively turned on and off in a timely manner based on low power consumption conditions. The echo signal from radar and the image data from vision are preprocessed respectively. Data gridding is used to map the object position information sensed by different sensors. The measured or estimated data (such as vehicle speed (moving speed or relative speed) or braking reaction time, object size, etc.) and data quality (reliability) are fused together. Alternatively, dynamic prediction of oncoming vehicles can be made based on data with high reliability.

[0068] 2. Judgment of the reliability of sensing data: Consider the data fusion process of radar confidence LR and visual confidence LV in different scenarios disclosed in this paper, as well as the processing when both visual confidence LV and radar confidence LR are too low.

[0069] 3. When too many moving objects are detected (e.g., the number of approaching vehicles is greater than 5, 10 or more), in order to balance low power consumption and improve computing efficiency, it is advisable to prioritize the identification of the top five (or other numerical) potential threat vehicles (but not limited to vehicle types (e.g., bicycles, motorcycles, cars, trucks, vans, etc.)). For example, the threat level of approaching vehicles can be determined based on the relationship between relative vehicle speed and braking reaction time / distance.

[0070] 4. The position of an object can be distinguished by different lanes and relative distances (far, near, etc.), and can be displayed (or shown) in the human-machine interface. The human-machine interface can adaptively provide warning signals in different situations.

[0071] In summary, although this disclosure has been presented above with reference to embodiments, it is not intended to limit this disclosure. Those skilled in the art to which this disclosure pertains can make various modifications and refinements without departing from the spirit and scope of this disclosure, and such modifications and refinements are not limited to the embodiments of this invention, but remain within the protection scope of this invention. Therefore, the protection scope of this disclosure shall be determined by the appended claims. [Simplified Explanation of the Diagram]

[0030] Figure 1 illustrates a functional block diagram of a traffic warning system according to an embodiment of the present invention. Figures 2A and 2B illustrate flowcharts of the operation method of the traffic warning system of Figure 1. Figure 3 illustrates a schematic diagram of object recognition in different lanes based on a visual algorithm disclosed herein. Figure 4 illustrates a schematic diagram of the relative relationship between the relative speed (horizontal axis) of an object relative to the traffic warning system and the stopping distance (vertical axis) corresponding to a specific relative speed. Figure 5 illustrates a schematic diagram of a collision area and a threat area in one embodiment. Figure 6 illustrates a schematic diagram of the relative relationship between the relative speed (horizontal axis) of an object relative to the traffic warning system and the corresponding stopping distance (vertical axis) according to another embodiment.

Claims

1. A traffic warning system, comprising: A vision sensor for detecting visual sensing data of at least one object; A radar sensor for detecting radar sensing data of each of the at least one object; a processor electrically connected to the vision sensor and the radar sensor, and configured to: obtain a radar confidence level of the radar sensing data by analyzing the radar sensing data; obtain a visual confidence level of the visual sensing data by analyzing the visual sensing data; if both the radar confidence level and the visual confidence level are greater than a preset value, obtain the position and velocity of the at least one object by analyzing the visual sensing data and the radar sensing data; and if the radar confidence level and the visual confidence level are not greater than the preset value and the visual confidence level is higher than the radar confidence level, determine the number of the at least one object approaching the traffic warning system based on the visual sensing data and the radar sensing data.

2. The traffic warning system as described in claim 1, wherein the preset value is between 70% and 90%.

3. The traffic warning system as described in claim 1, wherein the processor is configured to: determine the number of at least one object approaching the traffic warning system if both the radar confidence level and the visual confidence level are greater than a preset value.

4. The traffic warning system as described in claim 3, wherein the processor is configured to: output a warning signal if the number of the at least one object is not greater than a preset number.

5. The traffic warning system as described in claim 3, wherein the processor is configured to: if the number of the at least one object is greater than a preset number, output a warning signal for the object closest to the traffic warning system.

6. The traffic warning system as claimed in claim 1, wherein the processor is configured to: determine the number of at least one object approaching the traffic warning system based on the radar sensing data if the radar confidence level and the visual confidence level are not greater than the preset value and the radar confidence level is higher than the visual confidence level.

7. The traffic warning system as described in claim 6, wherein the processor is configured to: output a warning signal if the number of the at least one object is not greater than a preset number.

8. The traffic warning system as claimed in claim 6, wherein the processor is configured to: output a warning signal for the object closest to the traffic warning system if the number of the at least one object is greater than a preset number.

9. The traffic warning system as claimed in claim 1, wherein the processor is configured to: output a warning signal if the number of the at least one object is not greater than a preset number.

10. The traffic warning system as claimed in claim 1, wherein the processor is configured to: output a warning signal for the object closest to the traffic warning system if the number of the at least one object is greater than a preset number.

11. The traffic warning system as claimed in claim 1, wherein the processor is configured to: determine whether the at least one object is detected; and if no object is detected, keep the radar sensor on and keep the visual sensor off.

12. The traffic warning system as claimed in claim 1, wherein the processor is configured to: determine whether the at least one object is detected; and if the radar sensor detects the at least one object, activate the visual sensor.

13. The traffic warning system as described in claim 1, wherein the processor is configured to: periodically activate the visual sensor.

14. A method for operating a traffic warning system, comprising: A visual sensor detects visual sensing data of at least one object; a radar sensor detects radar sensing data of the at least one object; a processor obtains a radar confidence level of the radar sensing data by analyzing the radar sensing data; the processor obtains a visual confidence level of the visual sensing data by analyzing the visual sensing data; if both the radar confidence level and the visual confidence level are greater than a preset value, the processor obtains the position and speed of the at least one object by analyzing the visual sensing data and the radar sensing data; and if the radar confidence level and the visual confidence level are not greater than the preset value and the visual confidence level is higher than the radar confidence level, the processor determines the number of at least one objects approaching the traffic warning system based on the visual sensing data and the radar sensing data.

15. The operating method as described in claim 14, wherein the preset value is between 70% and 90%.

16. The method of operation as described in claim 14 further includes: If both the radar reliability and the visual reliability are greater than preset values, the number of at least one object approaching the traffic warning system is determined.

17. The operating method as described in claim 16 further includes: If the number of at least one object is not greater than a preset number, the processor outputs a warning signal.

18. The operating method as described in claim 16 further includes: If the number of at least one object is greater than a preset number, the processor outputs a warning signal for the object closest to the traffic warning system.

19. The operating method as described in claim 14, further comprising: If the number of at least one object is not greater than a preset number, the processor outputs a warning signal.

20. The operating method as described in claim 14 further includes: If the number of at least one object is greater than a preset number, the processor outputs a warning signal for the object closest to the traffic warning system.

21. The operating method as described in claim 14 further includes: The processor determines whether at least one object has been detected. If no object is detected, the processor keeps the radar sensor on and the vision sensor off.

22. The operating method as described in claim 14 further includes: The processor determines whether at least one object has been detected. And if the radar sensor detects the at least one object, the processor activates the visual sensor.

23. The operating method as described in claim 14 further includes: The vision sensor is turned on periodically.

24. A non-transitory computer-readable medium storing a plurality of instructions executed by at least one processor of a traffic warning system to cause the traffic warning system to perform the following methods: detecting visual sensing data of at least one object; detecting radar sensing data of at least one object; obtaining a radar confidence level of the radar sensing data by analyzing the radar sensing data; obtaining a visual confidence level of the visual sensing data by analyzing the visual sensing data; if both the radar confidence level and the visual confidence level are greater than a preset value, obtaining the position and speed of the at least one object by analyzing the visual sensing data and the radar sensing data; and if the radar confidence level and the visual confidence level are not greater than the preset value and the visual confidence level is higher than the radar confidence level, the processor determines the number of at least one objects approaching the traffic warning system based on the visual sensing data and the radar sensing data.

25. A non-transitory computer-readable medium storing the instructions as described in claim 24, the instructions being executed by the at least one processor of the traffic warning system to cause the traffic warning system to perform another method as described in any one of claims 15 to 23.

Images