WO2020211658A1

WO2020211658A1 - Trigger detection method, apparatus and system

Info

Publication number: WO2020211658A1
Application number: PCT/CN2020/083093
Authority: WO
Inventors: 李扬彦; 单单; 陈颖; 王振华
Original assignee: 阿里巴巴集团控股有限公司
Priority date: 2019-04-17
Filing date: 2020-04-03
Publication date: 2020-10-22
Also published as: CN111856611A

Abstract

Disclosed are a trigger detection method, apparatus and system. The system comprises a road surface detection apparatus for detecting, by means of a first sensor buried in a target detection area on a road surface, an event of a vehicle passing through the target detection area; a roadside detection apparatus for detecting, by means of a second sensor mounted on a roadside device and capable of covering the target detection area, the event of a vehicle passing through the target detection area; and a fusion algorithm processing apparatus for determining, according to the fusion of a road surface detection result from the road surface detection apparatus and a roadside detection result from the roadside detection apparatus, whether there is a vehicle passing through the target detection area, and generating a trigger signal according to a determination result. By means of the embodiments of the present application, the false detection rate and the missing detection rate in a trigger detection phase can be reduced, thereby improving the accuracy of license plate recognition and enabling senseless payment to be more fully applied in an expressway scenario.

Description

Trigger detection method, device and system

This application claims the priority of a Chinese patent application with an application number of 201910309995.2 and an invention title of "trigger detection method, device and system" filed on April 17, 2019, the entire content of which is incorporated into this application by reference.

Technical field

This application relates to the technical field of trigger detection, and in particular to trigger detection methods, devices and systems.

Background technique

Facing the increasing traffic pressure on expressways, more and more expressway operators have realized that introducing diversified payment methods can make up for the shortcomings of existing payment methods on high-level highways, thereby improving vehicle traffic efficiency and reducing the pressure on toll collectors. "Waitless payment" uses the Internet + electronic payment technology. Through the vehicle's "license plate payment" concept, the vehicle's license plate is used as a unique identifier to transform the cash collection method at the toll station into an online mobile payment method. Before use, the car owner can first bind his account registered in the mobile payment tool with the license plate. In addition, he can also enable functions such as "secret-free payment". The license plate number becomes a "payment code" and the vehicle is parked in and out At the toll station or highway toll station, the high-definition camera installed at the toll station automatically raises the pole and automatically deducts the fee after recognizing the license plate, without stopping the entire process. Compared with the ETC charging model, non-inductive payment is not only easy to open, but also does not require on-board OBU (Onboard Unit) equipment, and the payment method is also very convenient and flexible.

However, at present, non-inductive payment has been relatively fully applied only in the parking lot scene, and in the highway scene, it is subject to some restrictions. This is because, in the highway scene, the driving speed of the vehicle tends to be relatively fast, and the recognition accuracy of the existing license plate recognition technology in the highway scene has not yet reached the relevant requirements. At present, many suppliers claim that their license plate recognition rate has reached about 99%. However, in harsh environments, such as heavy rain, storms, haze, heavy snow, etc., most license plate recognition rates will drop by about 3% year-on-year. Therefore, in risk prevention and control On the highway, non-sense payment on expressways still needs to pass card auxiliary verification.

Therefore, how to improve the accuracy of license plate recognition so that non-inductive payment can be more fully applied in highway scenes has become a technical problem that needs to be solved by those skilled in the art.

Summary of the invention

This application provides a trigger detection method, device and system, which can reduce the false detection rate and the missed detection rate in the trigger detection stage, which is beneficial to improve the accuracy rate of license plate recognition, so that non-inductive payment can be more fully applied in the highway scene .

This application provides the following solutions:

A trigger detection system includes:

The road surface detection device is used to detect the event that a vehicle passes through the target detection area through the first sensor buried in the target detection area on the road surface;

The roadside detection device is used to detect the event that a vehicle passes through the target detection area through a second sensor installed on the roadside equipment that can cover the target detection area;

A fusion algorithm processing device for fusing the road surface detection result of the road surface detection device with the roadside detection result of the roadside detection device to determine whether there is a vehicle passing by the target detection area, and to generate a trigger signal according to the determination result .

A trigger detection method includes:

Receive the road surface detection result of the road surface detection device and the roadside detection result of the roadside detection device, wherein the road surface detection device is used to detect the detection of the vehicle passing the target detection area through the first sensor buried in the target detection area on the road surface Event detection, and the roadside detection device is configured to detect an event when a vehicle passes through the target detection area through a second sensor installed on a roadside device that can cover the target detection area;

By fusing the road surface detection result with the roadside detection result, judging whether there is a vehicle passing by the target detection area;

Generate a trigger signal according to the judgment result.

A trigger detection method includes:

Generate a trigger signal according to the judgment result.

A trigger detection system includes:

A visual detection device, which is used for a visual sensor installed on roadside equipment that can cover the target detection area, and detects the event of a vehicle passing through the target detection area;

A radar detection device, which is used to detect an event that a vehicle passes through the target detection area through a radar sensor installed on the roadside equipment and capable of covering the target detection area;

The fusion algorithm processing device is used for fusing the visual detection result of the vision detection device with the radar detection result of the radar detection device, judging whether there is a vehicle passing by the target detection area, and generating a trigger signal according to the judgment result.

A trigger detection method includes:

Receive the visual detection result of the visual detection device and the radar detection result of the radar detection device, wherein the visual detection device is used to detect the vehicle passing the target through the visual sensor installed on the roadside equipment and covering the target detection area Detection of regional events, and the radar detection device is used to detect the event of vehicles passing through the target detection area through a radar sensor installed on roadside equipment and capable of covering the target detection area;

By fusing the visual detection result with the radar detection result, judging whether there is a vehicle passing by the target detection area;

Generate a trigger signal according to the judgment result.

A trigger detection device includes:

The first result receiving unit is used to receive the road surface detection result of the road surface detection device and the roadside detection result of the roadside detection device, wherein the road surface detection device is used to pass the first sensor buried in the target detection area on the road surface The vehicle passing through the target detection area is used to detect the event of the vehicle passing the target detection area, and the roadside detection device is used to detect the event of the vehicle passing the target detection area through a second sensor installed on the roadside equipment and capable of covering the target detection area. Carry out testing;

The first fusion calculation unit is configured to determine whether a vehicle passes by the target detection area by fusing the road surface detection result with the roadside detection result;

The first signal generating unit is configured to generate a trigger signal according to the judgment result.

A trigger detection device includes:

The second result receiving unit is used to receive the visual detection result of the visual detection device and the radar detection result of the radar detection device, wherein the visual detection device is used to pass a vision sensor installed on the roadside equipment that can cover the target detection area , Detecting the event of the vehicle passing the target detection area, and the radar detection device is used to detect the event of the vehicle passing the target detection area through the radar sensor installed on the roadside equipment and covering the target detection area. Carry out testing;

The second fusion calculation unit is configured to determine whether a vehicle passes by the target detection area by fusing the visual detection result with the radar detection result;

The second signal generating unit is used to generate a trigger signal according to the judgment result.

A roadside device, including:

The roadside detection device is used to detect the event that a vehicle passes through the target detection area through a second sensor installed on the roadside equipment and capable of covering the target detection area;

The fusion algorithm processing device is used for fusing the road surface detection result of the road surface detection device with the roadside detection result of the roadside detection device, judging whether there is a vehicle passing by the target detection area, and generating a trigger signal according to the judgment result; wherein The road surface detection device is used to detect the event that the vehicle passes through the target detection area through the first sensor buried in the target detection area on the road surface.

A roadside device, including:

According to the specific embodiments provided in this application, this application discloses the following technical effects:

Through the embodiments of the present application, the detection results of the event that the road detection device and the roadside detection device pass the target detection area can be merged to determine whether there is a vehicle passing by the target detection area, and a trigger signal is generated according to the judgment result to Used to trigger operations such as license plate recognition. Wherein, the road surface detection device can detect the event of the vehicle passing the target detection area through the first sensor buried in the target detection area on the road surface, and the roadside detection device can be installed on the roadside equipment to cover the target The second sensor in the detection area detects the event that the vehicle passes through the target detection area. In this way, since the detection results of the event that the vehicle enters the target detection area can be triggered by fusing the two systems, the false detection rate and the missed detection rate can be reduced. In addition, due to the different working principles of the two systems and the large distance between the installation positions (one buried under the road surface and one installed on the roadside equipment), they have almost completely different failure modes, which greatly improves the overall triggering The mean time between failures of the detection system. Furthermore, both systems have lower hardware costs, especially for sensors installed on roadside equipment, which are shared hardware equipment with functions such as vehicle-road coordination. For road sections where roadside equipment has been installed, it is even more important. There is no need to add new hardware costs. Therefore, the overall hardware cost of the trigger detection system in the embodiment of the present application is lower, which is conducive to more full application of non-inductive payment in highway scenarios.

In another embodiment, the trigger detection may be performed by specifically sensing the mutual fusion between different sensors in the base station. In this way, because different types of sensors also have different working principles, although the installation locations are relatively close, there will be obvious differences in failure modes. In addition, this method does not need to lay hardware equipment on the ground, so the requirements for road construction will be relatively low, and the sensor equipment used belongs to the equipment shared with functions such as vehicle-road coordination. Therefore, on the whole, Can also play a role in reducing hardware costs.

Of course, implementing any product of this application does not necessarily need to achieve all the advantages described above at the same time.

Description of the drawings

In order to explain the embodiments of the present application or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some of the present application. Embodiments, for those of ordinary skill in the art, without creative work, other drawings can be obtained from these drawings.

Figure 1 is a schematic diagram of a system provided by an embodiment of the present application;

Figure 2 is a flowchart of the first method provided by an embodiment of the present application;

Figure 3 is a schematic diagram of another system provided by an embodiment of the present application;

Figure 4 is a flowchart of a second method provided by an embodiment of the present application;

Figure 5 is a schematic diagram of a device provided by an embodiment of the present application.

detailed description

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art fall within the protection scope of this application.

First of all, it needs to be explained that in the process of license plate recognition, two stages are usually required. One is the trigger detection when the vehicle just enters the detection area, and the second is the continuous detection before the subsequent driving until it leaves the detection area. Among them, a license plate recognition device can be installed in the detection area, and the device can perform operations such as photographing and recognition under the trigger of specific instructions. As for the accuracy of license plate recognition, in addition to the accuracy of the specific recognition algorithm, it is also closely related to the accuracy of the specific trigger detection. In other words, an important factor that affects the rate of license plate recognition is the exposure timing of the camera that takes the license plate, and the selection of the exposure timing is based on the result of vehicle trigger detection. If you fail to find when a vehicle is passing by and fail to send a trigger command to the recognition device in time, it may cause the recognition device to fail to collect relevant license plate information in time, and then there is no way to recognize the license plate information. Therefore, the accuracy of the trigger detection phase is also crucial. In the embodiments of the present application, corresponding solutions are mainly provided for the trigger detection phase. Specifically, the so-called trigger detection is to detect the event that a vehicle enters the detection area, and use this as a basis to issue a trigger instruction to a specific license plate recognition device.

Among them, there are many specific implementation methods for detecting the event of a vehicle entering the detection area. For example, in one method, a road trigger detection system can be used to achieve this, for example, it can be implemented by an induction coil or a pressure gauge. Way to perform trigger detection. Among them, the induction coil is a ring-shaped sensor buried in the ground. When a large metal object (such as a vehicle) passes over the coil, the inductance value of the coil will change greatly, so the vehicle can be detected. The pressure gauge is a linear sensor array buried on the ground. When a heavier object (such as a vehicle) passes over the array, the voltage or resistance of the sensor array will change significantly, so the vehicle can be detected. However, the disadvantage of this road surface trigger detection system is that in the highway scene, the speed of the vehicle may be relatively high. In this case, the vehicle may quickly pass by a specific induction coil or pressure gauge. At this time, the probability of missed detection may be relatively high. In addition, this kind of detection system needs to modify the road surface, which is difficult to construct, and there may be many locations where toll points need to be set up on the expressway, for example, including high-speed entrances and exits, provincial boundaries, etc. Therefore, it is also possible It will affect normal road traffic.

In addition, there are some trigger detection schemes that can be considered, for example, by installing ultrasonic or infrared grating equipment at specific toll points, etc., through this equipment to detect the event of a vehicle entering the detection area. Although this method is less affected by the speed of the vehicle, it still needs to involve the construction and transformation of the road.

Based on the above analysis, the embodiment of the present application provides a new trigger detection method. In this method, the trigger detection function of the vehicle can be combined with the function of the sensing base station in the vehicle-road coordination system. Among them, in the vehicle-road coordination system, the sensing base station usually corresponds to the roadside equipment, where the full name of the roadside equipment can be the intelligent roadside equipment, which is the infrastructure located in the road subsystem. It mainly includes RSU (Roadside Unit, roadside unit, V2X equipment deployed on or near the road, including but not limited to roadside unit and vehicle-mounted unit communication (I2V), roadside unit and roadside unit communication (I2I) And other functions), physical units such as roadside sensors and computing storage units. Among them, the sensing methods of roadside sensors mainly include acquiring data from cameras, millimeter wave radars, lidars and other sensors, and processing the data through image or signal processing algorithms in the computing unit, and finally acquiring information on roads and traffic participants. For example, lane line coordinates, vehicle and pedestrian position, speed, etc. The vehicle can obtain information about the surrounding environment by receiving the above-mentioned broadcast message, and then combine its own speed, etc., to make driving decisions or realize driving assistance.

Therefore, in the embodiments of the present application, based on the roadside equipment in the above-mentioned vehicle-road coordination system, specific trigger detection can be realized through the integration of multiple trigger detection systems, and mutual verification between different trigger detection systems , To determine whether there is a vehicle in the specific detection area to detect this event, and then send a trigger instruction to the specific license plate recognition device. The fusion of such multiple trigger detection systems can reduce the false detection rate and missed detection rate of trigger detection, and provide a basis for improving the accuracy of license plate recognition.

Among them, for different specific scenarios, different fusion methods can also be adopted. For example, for a target location where specific vehicles are generally slower and have non-inductive billing requirements, such as the entrance or exit of an expressway, because the vehicle usually has a ramp section at such entrance or exit, the vehicle The driving speed will be relatively slow, which is conducive to the high reliability of the road trigger detection system. Therefore, in view of the above situation, the road surface trigger detection system can be combined with the vision or radar (millimeter wave radar or lidar, etc.) sensors in the roadside equipment, that is, the perception results of the vision, radar and other sensors in the base station can be sensed. It can also be added to the trigger detection, and the detection results of the two systems are fused and calculated to determine whether a specific vehicle is passing the detection area. In this way, when a vehicle is detected by either of the two systems, it can be detected by the hybrid trigger detection system, so the missed detection rate is greatly reduced. In addition, since the working principles of the two systems are completely different and the distance between the installation positions is large (one is buried under the road surface and the other is installed on the roadside equipment), they have almost completely different failure modes, which greatly improves the overall Trigger the mean time between failures of the detection system. Furthermore, both systems have lower hardware costs (for sensors installed on roadside equipment, they share hardware equipment with functions such as vehicle-road coordination, so the hardware cost will be lower), so the hybrid trigger detection The overall hardware cost of the system is low.

In addition, the main roads of expressways may also have the requirement of non-inductive charging. For example, when the vehicle crosses the boundary of different jurisdictions and other locations, segmented non-inductive charging may also be carried out, etc. Wait. In this scenario, because it is a main road and there are no road sections such as ramps on the speed limit, the speed of the vehicle may be relatively fast. Therefore, it is no longer suitable to use a road trigger detection system for trigger detection. In the sensing base station, since multiple different types of sensors, such as vision sensors, radar sensors, etc., are usually included, trigger detection can also be performed by specifically sensing the mutual fusion of different sensors in the base station. In this way, because different types of sensors also have different working principles, although the installation positions are relatively close, there will be obvious differences in failure modes. In addition, this method does not need to lay hardware equipment on the ground, so the requirements for road construction will be relatively low, and the sensor equipment used belongs to the equipment shared with functions such as vehicle-road coordination. Therefore, on the whole, Can also play a role in reducing hardware costs.

That is to say, in the embodiments of the present application, the sensors on the roadside equipment can be used as part of the data source to trigger detection, and the roadside equipment and various sensors set therein generally belong to the hardware equipment in the vehicle-road coordination system. In other words, during the design of the vehicle-road coordination system, it is also necessary to install and deploy the above-mentioned equipment. Therefore, when using the sensors on the roadside equipment for event sensing, at the hardware level, it only needs to share the same with the vehicle-road coordination system. A set of hardware equipment, and at the software level, you can additionally implement algorithms for event awareness and fusion algorithms. It can be seen that the overall hardware cost is under control, and the reconstruction of the road only involves the entrance or exit of the expressway, and the construction volume and difficulty are greatly reduced. In addition, the modification of the entrances and exits of the expressway does not affect the normal driving vehicles on the expressway, so the impact on the traffic efficiency is small. Of course, at the entrance or exit of the highway, a combination of visual sensors and radar sensors on the roadside equipment can also be used for trigger detection. In this case, lower hardware costs can be achieved.

The specific implementation scheme is described in detail below.

Example one

In the first embodiment, first, the solution of fusing the road surface trigger detection and the roadside device trigger detection is introduced. Specifically, referring to Fig. 1, the first embodiment of the present application first provides a trigger detection system, which may include:

The road surface detection device 101 is configured to detect the event of a vehicle passing the target detection area through the first sensor buried in the target detection area on the road surface;

The roadside detection device 102 is configured to detect an event that a vehicle passes through the target detection area through a second sensor installed on a roadside device that can cover the target detection area;

The fusion algorithm processing device 103 is used for fusing the road surface detection result of the road surface detection device with the roadside detection result of the roadside detection device, judging whether there is a vehicle passing by the target detection area, and generating a trigger according to the judgment result signal.

Among them, the requirement for the first sensor buried on the road surface is that whenever a vehicle passes by the first sensor, at least one physical quantity of the first sensor undergoes a rapid, strong and predictable change, which overcomes the speed of the vehicle and the speed of the vehicle. The size and weight vary widely, and problems such as bumps occur during the driving of the vehicle. In specific implementation, the first sensor may include an induction coil or a pressure gauge. Among them, if an induction coil is used, the first sensor only has a vehicle trigger detection function, which is low in cost; if a pressure gauge is used, it can trigger detection of the vehicle and also obtain additional information such as vehicle speed and wheelbase, but the cost is slightly higher. Therefore, the former can be referred to as "low-distribution system", and correspondingly, the latter as "high-distribution system".

Specifically, the low-profile system may include induction coils, feeders, and detectors. When a large metal object (such as a vehicle) passes by the coil, the inductance value of the coil decreases, and the frequency value of the output signal of the oscillating circuit in the detector becomes larger; when the change exceeds a certain threshold and When detected by the frequency detection circuit, the detector considers that there is currently a vehicle passing by and outputs the detection result, otherwise it defaults to no vehicle passing by.

The high-profile system can include pressure gauges, feeders and detectors. When the vehicle passes the detection device, the wheels of each axle roll over multiple linear sensors in turn, so that their voltages (or resistance values) change sequentially. These changes are adjusted to the appropriate dynamic range by the multi-channel signal amplifier, and then sent to the multi-channel analog-to-digital converter for simultaneous sampling. The sampling result is sent to the signal processing algorithm as a digital signal, and finally whether there is a vehicle passing, the direction of movement, Test results of vehicle speed and wheelbase.

The second sensor may specifically include a vision sensor or a radar sensor. Among them, the visual trigger detection system can detect vehicles on the road in real time through artificial intelligence algorithms and computing capabilities, and can distinguish the lane where the vehicle is located. The visual detection module takes the video image stream as input, and checks whether there is a vehicle in the current image frame of the video stream according to a certain fixed frequency. If there is a car, the system will notify the relevant module and output the vehicle position to it.

The radar trigger detection system can detect the changes in road clearance or electromagnetic wave frequency caused by vehicles passing by through millimeter wave radars installed above the road, and trigger detection of vehicles. Among them, millimeter wave radar is an active sensor based on electromagnetic waves. It transmits modulated electromagnetic waves in the millimeter wave frequency band in different directions and receives reflected signals, measures the distance and speed of objects that reflect the signals, describes these objects as targets and outputs them as detection results. The requirement for the millimeter-wave radar installed above the road is that whenever a vehicle passes through the monitoring system, at least one physical quantity measured by the radar undergoes a rapid, strong and predictable change, which overcomes the speed of the vehicle and the size of the vehicle. Large, bumps and other problems may occur during the driving of the vehicle.

The radar trigger detection system can include millimeter wave radar and signal receiver. The millimeter wave radar transmits a millimeter wave signal with a narrow beam and a known frequency to scan the road surface and measure the road clearance and Doppler shift in real time. When the vehicle passes the detection system, the road headroom will decrease and the Doppler frequency shift will increase, and the millimeter wave radar will detect the vehicle accordingly. Millimeter wave radar can detect multiple vehicles at the same time, and describe their position, speed, reflected signal strength and other information as multiple targets as the result output. This output is called multi-target information. The multi-target information is sent to the signal receiver, and after the false targets are filtered out through algorithm processing, the final detection result is generated.

Among them, the signal receiver receives the multi-target information output by the millimeter-wave radar in real time, and filters out false targets through algorithms. The main steps of the algorithm are to first read the map data and calibration parameters to obtain the movable area near the radar installation location, and filter out targets outside the movable area; secondly, for the targets in the movable area, you can check their reflection The signal strength is compared with the pre-stored statistical law to filter out targets that do not comply with the statistical law; the remaining targets are output as the final detection result.

After the above-mentioned road surface detection device and roadside detection device detect the event that the vehicle enters the target detection area, the detection result can be provided to the fusion algorithm processing device, and the system performs fusion processing to obtain the final judgment result. Among them, the fusion algorithm processing device can be specifically configured in the processor device on the roadside device, or can also be configured in other devices, which can be connected to the road surface detection device and the roadside detection device respectively. And the processor devices, etc. are in a static state, and the distances between them are relatively close. Therefore, the data connection can be realized in various ways such as wired or wireless, so that data can be transmitted between each other.

It should be noted here that the full name of the roadside device described in the embodiments of the present application may be an intelligent roadside device, which is an infrastructure located in a road subsystem. It mainly includes physical units such as RSU (roadside unit), roadside sensors, and calculation storage unit; among them, the calculation storage unit can be composed of two sub-units: intelligent calculation and storage. Different physical units can be deployed on different physical devices or in the same physical device. Therefore, in the specific implementation, the specific converged computing system can be deployed in the roadside equipment with the help of the above-mentioned characteristics of the roadside equipment, and it is only necessary to expand the dedicated intelligent computing and computing system based on the original foundation. Just store the sub-unit. In this way, the original intelligent calculation and storage subunit can still be used for related calculations in the vehicle-road coordination system, and the expanded intelligent calculation and storage subunit is used for the fusion calculation processing of trigger detection in the embodiment of this application .

Among them, when specifically performing fusion processing, the same vehicle corresponding to the road surface detection result and the roadside detection result can be identified first. In specific implementation, since the first sensor is just below the ground of the target detection area, its detection result naturally has the attribute of location information, and although the second sensor is installed on the roadside equipment, it also has the ability to sense location, so , The position information it senses can be carried in the roadside detection result. In this way, the fusion algorithm processing device can make a judgment after receiving the detection results of the two, if the road detection result indicates that a vehicle has passed the target detection area, and the roadside detection result indicates that the vehicle's location information is located at the Within the target detection area, the vehicle included in the road surface detection result and the vehicle in the roadside detection result are determined to be the same vehicle, and it is determined that the vehicle is passing through the target detection area.

In addition, in specific implementation, the road surface detection result and the roadside detection result can directly exist in two states of "Yes" and "No". If the detection results of both at a certain moment are both "Yes", the output will be output after fusion. Yes, if both are "No", then output "No" after fusion. However, if one is "yes" and the other is "no", it means that the detection results of the two are contradictory. Therefore, there may be situations such as failure to be fused, which does not conform to the original intention of the fusion algorithm. Therefore, in an optional implementation manner, in order to further improve the reliability of trigger detection, a compromise solution may also be adopted to fuse the two detection results. Specifically, after receiving the detection results of the two detection systems, in the process of fusion calculation, the probability calculation can be performed first, and then the fusion can be performed to obtain the result of the fusion calculation. That is to say, after receiving the detection results of the two detection systems, first determine the respective "confidence" of the two, and then add the two, and compare the result with the decision threshold to improve The reliability of specific fusion results.

Among them, when calculating the probability, the two detection systems can first calculate the probability in real time according to the detection algorithm. Specifically, this probability is called the posterior probability. From the mathematical principle, it can be calculated from the detection result and the template signal or training The degree of matching between the models is determined. In this way, when performing fusion calculation, you can proceed as follows:

y=p1×r1+p2×r2

Among them, p1 and p2 are the weights obtained through the posterior probability; r1 and r2 are the initial detection results, 1 means that there is a car in the designated area, and 0 means that there is no car.

After that, the fusion algorithm compares y with the decision threshold to get the final result of whether there is a car passing by.

In specific implementation, the calculation of the posterior probability can be completed within each detection system. Specifically, the road detection device can determine the first posterior probability according to the matching degree between the initial detection result and the template signal/training model; the roadside detection device can determine the first posterior probability according to the matching between the initial detection result and the template signal/training model Degree, determine the second posterior probability. In this way, the first initial detection result can be multiplied by the first posterior probability to obtain the first probability; the second initial detection result can be multiplied by the second posterior probability to obtain the second probability. In this way, after the two probabilities received by the fusion calculation system are added, they can be used as the result of the fusion calculation, and then the result is compared with the decision threshold to get the final result of whether there is a car passing by. That is to say, for each detection system, the initial detection result is 0 or 1. After obtaining this initial detection result, it is not directly sent to the fusion computing system for fusion calculation, but can be based on The matching degree with the template signal/training model is calculated, the posterior probability is calculated, and the product is multiplied with the initial detection result, and the calculated probability value is used as the actual output value of the detection system. In this way, the probability value of the data received by the fusion computing system is no longer the initial detection result of 0 or 1.

Or, in order to obtain the fusion detection result more accurately, a prior probability component can also be introduced. Specifically, because the confidence of the detection result obtained by the specific detection system may be affected by factors such as weather, vehicle speed, etc., this influence can be obtained through the analysis of some historical data or machine learning. Therefore, it can be provided based on the pre-stored statistical results, for example, different weather and vehicle speeds correspond to different detection probabilities to obtain the prior probability. Among them, for different detection systems, due to their different working principles, etc., even under the same weather, vehicle speed and other conditions, the confidence of the detection results may be different. Therefore, the respective detection systems can be obtained separately. The prior probability. In this way, by combining these two probabilities, the problem of the accuracy of the fusion detection result can be solved more accurately. For example, in one of the implementation methods, the mathematical description of the specific algorithm can be as follows:

y=c1×x1+c2×x2,

Among them, c1 and c2 are the weights obtained by the prior probability, x1 and x2 are the outputs of the two detection systems, and y is the output of the fusion algorithm. For x1 and x2, their calculation methods are as follows:

x1=p1×r1, x2=p2×r2

Finally, the fusion algorithm compares y with the decision threshold to get the final result of whether there is a car passing by.

In specific implementation, the introduction of a priori probability can be completed by the fusion computing system, that is, after receiving the road detection result and the roadside detection result (which has been calculated based on the posterior probability), the The historical statistical results corresponding to the respective detection systems determine the first prior probability corresponding to the road surface detection device and the second prior probability corresponding to the roadside detection result. Then, the road surface detection result can be multiplied by the first prior probability, the roadside detection result can be multiplied by the second prior probability, and the products of the two can be added together to obtain the final fusion calculation result. After that, the fusion calculation result is compared with the judgment threshold to determine the fusion judgment result, that is, whether there is a vehicle passing by in the target area.

Wherein, in specific implementation, after the fusion algorithm processing device generates a specific trigger signal according to the specific fusion estimation result, the trigger signal may also be sent to the license plate recognition device, and the license plate recognition device is used to recognize the license plate of the vehicle , In order to charge for the vehicle.

Among them, under the solution provided in the first embodiment of the present application, the specific license plate recognition device, the first sensor, and the roadside device can be set at a target location on the expressway that has the characteristics of slow driving of the vehicle and the demand for non-inductive charging Place. For example, specifically, it may be set at the entrance/exit of the expressway to perform non-inductive charging processing for vehicles entering or leaving the expressway, and so on.

It should be noted that, in specific implementation, the fusion estimation result provided by the embodiment of the present application may also be provided to a processing module for monitoring traffic flow, and so on.

In short, through the embodiments of the present application, the detection results of the road surface detection device and the roadside detection device vehicle passing the target detection area can be merged to determine whether there is a vehicle passing by the target detection area, and a trigger signal is generated according to the judgment result , To trigger operations such as license plate recognition. Wherein, the road surface detection device can detect the event of the vehicle passing the target detection area through the first sensor buried in the target detection area on the road surface, and the roadside detection device can be installed on the roadside equipment to cover the target The second sensor in the detection area detects the event that the vehicle passes through the target detection area. In this way, since the detection results of the event that the vehicle enters the target detection area can be triggered by fusing the two systems, the false detection rate and the missed detection rate can be reduced. In addition, due to the different working principles of the two systems and the large distance between the installation positions (one is buried under the road surface and the other is suspended on the roadside equipment), they have almost completely different failure modes, which greatly improves the overall Trigger the mean time between failures of the detection system. Furthermore, both systems have lower hardware costs, especially for sensors installed on roadside equipment, which are shared hardware equipment with functions such as vehicle-road coordination. For road sections where roadside equipment has been installed, it is even more important. There is no need to add new hardware costs, so the overall hardware cost of the trigger detection system in the embodiment of the present application is lower.

Example two

The second embodiment corresponds to the first embodiment. From the perspective of a specific fusion algorithm processing device, a trigger detection method is provided. Referring to FIG. 2, the method may specifically include:

S201: Receive the road surface detection result of the road surface detection device and the roadside detection result of the roadside detection device, wherein the road surface detection device is used to detect the vehicle passing the target through the first sensor buried in the target detection area on the road surface Detecting an event in the area, and the roadside detection device is used to detect an event in which a vehicle passes through the target detection area through a second sensor installed on a roadside device that can cover the target detection area;

S202: Determine whether a vehicle passes by the target detection area by fusing the road surface detection result with the roadside detection result;

S203: Generate a trigger detection signal according to the judgment result.

Wherein, the roadside detection result also includes the position information of the detected vehicle; at this time, specifically when judging whether there is a vehicle passing by the target detection area, the following can be performed: if the roadside detection result indicates that there is a vehicle passing by In the target detection area, and the roadside detection result indicates that the position information of the vehicle is within the target detection area, the vehicle included in the road surface detection result and the vehicle in the roadside detection result are determined to be the same vehicle, And it is determined that the vehicle is passing through the target detection area.

In addition, during specific implementation, the road surface detection result may include: a first probability of a vehicle passing the target detection area; the roadside detection result may include: a second probability of a vehicle passing the target detection area; When judging whether there is a vehicle passing by the target detection area, it may be determined whether there is a vehicle passing by the target detection area according to the first probability, the second probability and the judgment threshold information in the preset estimation algorithm.

Specifically, the first probability is determined by the road surface detection device according to the initial detection result and the posterior probability corresponding to the road surface detection device;

The second probability is determined by the roadside detection device according to the initial detection result and the posterior probability corresponding to the roadside detection device;

Wherein, the posterior probability is determined according to the degree of matching between the initial detection result and the template signal and/or the pre-trained model;

The initial detection result is: the respective detection system's perception result of whether there is a vehicle passing by the target detection area.

Or, further, during the fusion calculation, the first prior probability corresponding to the road surface detection device may be determined based on the current weather and/or vehicle speed factors and the pre-stored historical statistical results corresponding to the respective detection systems , And the second prior probability corresponding to the roadside detection device; the historical statistical result includes: the confidence level of the initial detection result of the respective detection system under the corresponding weather and/or vehicle speed condition;

Then, the product of the first probability and the first prior probability and the product of the second probability and the second prior probability are added, and the result of the addition is compared with the preset threshold information In this way, it is determined whether there is a vehicle passing by the target detection area.

Example three

The third embodiment provides another trigger detection system. The trigger detection system can also include two sets of trigger detection systems, which can be implemented by two different sensor devices installed on the roadside equipment that can cover the target detection area. Detect and merge the detection results to improve the reliability of specific trigger detection. Specifically, referring to Figure 3, the system may include:

The vision detection device 301 is used for a vision sensor that is installed on roadside equipment and can cover the target detection area, and detects the event of a vehicle passing the target detection area;

The radar detection device 302 is configured to detect an event that a vehicle passes through the target detection area through a radar sensor installed on the roadside equipment and capable of covering the target detection area;

The fusion algorithm processing device 303 is used for fusing the visual detection result of the visual detection device with the radar detection result of the radar detection device, judging whether there is a vehicle passing by the target detection area, and generating a trigger detection signal according to the judgment result .

The fusion algorithm processing device may also be used to send the trigger detection signal to a license plate recognition device, and the license plate recognition device is used to recognize the license plate of a vehicle so as to perform charging processing on the vehicle.

In the third embodiment, the license plate recognition device and the roadside device are more suitable to be installed at a target location on an expressway that has the characteristics of fast driving of vehicles and has a demand for non-inductive charging. For example, it can be specifically at the boundary of the jurisdiction on an expressway, which is usually located on the main road of the expressway, and the speed of the vehicle is usually relatively fast. Therefore, the third embodiment can be used. In order to realize the non-sense charging processing for vehicles entering different jurisdictions. Of course, it is also feasible if it is installed at the entrance or exit of the expressway.

Among them, the specific visual trigger detection system and radar trigger detection system in the third embodiment may be the same as those in the first embodiment, and will not be described in detail here.

Through this third embodiment, it is possible to perform trigger detection by specifically sensing the mutual fusion between different sensors in the base station. In this way, because different types of sensors also have different working principles, although the installation positions are relatively close, there will be obvious differences in failure modes. In addition, this method does not need to lay hardware equipment on the ground, so the requirements for road construction will be relatively low, and the sensor equipment used belongs to the equipment shared with functions such as vehicle-road coordination. Therefore, on the whole, Can also play a role in reducing hardware costs.

Example four

The fourth embodiment corresponds to the first embodiment. From the perspective of a specific fusion algorithm processing device, a trigger detection method is provided. Referring to FIG. 4, the method may specifically include:

S401: Receive the visual detection result of the vision detection device and the radar detection result of the radar detection device, where the vision detection device is used to pass the vehicle through the visual sensor installed on the roadside equipment and covering the target detection area. Detecting an event in a target detection area, and the radar detection device is used to detect an event in which a vehicle passes through the target detection area through a radar sensor installed on a roadside device that can cover the target detection area;

S402: Determine whether there is a vehicle passing by the target detection area by fusing the visual detection result with the radar detection result;

S403: Generate a trigger detection signal according to the judgment result.

In specific implementation, the visual detection result may include: a third probability of a vehicle passing the target detection area; the radar detection result may include: a fourth probability of a vehicle passing the target detection area; in this case, According to the third probability, the fourth probability and the judgment threshold information in the preset estimation algorithm, it is judged whether there is a vehicle passing by the target detection area.

For the parts that are not described in detail in the foregoing Embodiment 2 to Embodiment 4, reference may be made to the record in the foregoing Embodiment 1, which will not be repeated here.

Corresponding to the second embodiment, the embodiment of the present application also provides a trigger detection device. Referring to FIG. 5, the device may specifically include:

The result receiving unit 501 is configured to receive the road surface detection result of the road surface detection device and the roadside detection result of the roadside detection device, wherein the road surface detection device is used to detect the vehicle through the first sensor buried in the target detection area on the road surface The event that passes through the target detection area is detected, and the roadside detection device is used to detect the event that the vehicle passes through the target detection area through a second sensor installed on the roadside equipment that can cover the target detection area. Detection

The fusion calculation unit 502 is configured to determine whether a vehicle passes by the target detection area by fusing the road surface detection result with the roadside detection result;

The signal generating unit 503 is configured to generate a trigger signal according to the judgment result.

Wherein, the roadside detection result also includes the position information of the detected vehicle;

The fusion calculation unit may be specifically used for:

If the road surface detection result indicates that a vehicle passes through the target detection area, and the roadside detection result indicates that the position information of the vehicle is within the target detection area, then the vehicle and the roadside included in the road surface detection result The vehicle in the detection result is determined to be the same vehicle, and it is determined that the vehicle is passing through the target detection area.

In specific implementation, the road surface detection result includes: a first probability of a vehicle passing through the target detection area;

The roadside detection result includes: a second probability of a vehicle passing through the target detection area;

The fusion calculation unit may be specifically used for:

According to the first probability, the second probability and the judgment threshold information in the preset estimation algorithm, it is judged whether there is a vehicle passing by the target detection area.

Wherein, the first probability is determined by the road surface detection device according to the initial detection result and the posterior probability corresponding to the road surface detection device;

Wherein, the posterior probability is determined according to the degree of matching between the initial detection result and the template signal/training model;

During specific implementation, the first fusion computing unit may specifically include:

The priori probability determination subunit is used to determine the first priori probability corresponding to the road surface detection device and the roadside according to the current weather and/or vehicle speed factors and the pre-stored historical statistical results corresponding to the respective detection systems The second prior probability corresponding to the detection device; the historical statistical result includes: the confidence level of the initial detection result of the respective detection system under the corresponding weather and/or vehicle speed condition;

The judging subunit is used to add the product of the first probability and the first prior probability, and the product of the second probability and the second prior probability, and by adding the result of the addition to preset threshold information The way of comparison is to determine whether a vehicle passes by the target detection area.

Corresponding to the fourth embodiment, the embodiment of the present application also provides a trigger detection device. Referring to FIG. 5, the device may specifically include:

The result receiving unit is used to receive the visual detection result of the visual detection device and the radar detection result of the radar detection device, wherein the visual detection device is used to pass the visual sensor installed on the roadside equipment that can cover the target detection area. The vehicle passes through the target detection area to detect the event, and the radar detection device is used to detect the vehicle passes through the target detection area through the radar sensor installed on the roadside equipment and can cover the target detection area ；

The fusion calculation unit is used to determine whether a vehicle passes by the target detection area by fusing the visual detection result with the radar detection result;

The signal generating unit is used to generate a trigger signal according to the judgment result.

In specific implementation, the visual detection result includes: a third probability that a vehicle passes through the target detection area;

The radar detection result includes: a fourth probability that a vehicle passes through the target detection area;

The fusion calculation unit may be specifically used for:

According to the third probability, the fourth probability and the judgment threshold information in the preset estimation algorithm, it is judged whether there is a vehicle passing by the target detection area.

Example five

As mentioned above, in the embodiments of this application, the specific fusion algorithm processing device can be deployed in multiple specific devices. In one of the solutions, since the roadside device has its own computing capabilities and data storage capabilities, therefore, It is also possible to directly implement the aforementioned fusion algorithm processing device in the roadside equipment. Specifically, Embodiment 5 of the present application also provides a roadside device, which may specifically include:

Example Six

The solution of the fifth embodiment above corresponds to the first and second embodiments, and among the solutions corresponding to the third and fourth embodiments, the embodiment of the present application also provides another roadside device, including:

Regarding the details in the fifth and sixth embodiments above, please refer to the records in the foregoing embodiments, which will not be repeated here.

From the description of the foregoing implementation manners, it can be understood that those skilled in the art can clearly understand that this application can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution of this application essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as ROM/RAM, magnetic disk , CD-ROM, etc., including several instructions to enable a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in each embodiment or some parts of the embodiment of this application.

The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment. The system and system embodiments described above are merely illustrative, where the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, namely It can be located in one place, or it can be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the objectives of the solutions of the embodiments. Those of ordinary skill in the art can understand and implement it without creative work.

The trigger detection method, device and system provided by this application are described in detail above. Specific examples are used in this article to illustrate the principles and implementation of this application. The description of the above embodiments is only used to help understand the application. The method and its core idea; meanwhile, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation and the scope of application. In summary, the content of this specification should not be construed as a limitation on this application.

Claims

A trigger detection system is characterized by comprising:

The road surface detection device is used to detect the event that a vehicle passes through the target detection area through the first sensor buried in the target detection area on the road surface;

The roadside detection device is used to detect the event that a vehicle passes through the target detection area through a second sensor installed on the roadside equipment that can cover the target detection area;

A fusion algorithm processing device for fusing the road surface detection result of the road surface detection device with the roadside detection result of the roadside detection device to determine whether there is a vehicle passing by the target detection area, and to generate a trigger signal according to the determination result .
The system according to claim 1, wherein:

The fusion algorithm processing device is also used to send the trigger signal to a license plate recognition device, and the license plate recognition device is used to recognize the license plate of the vehicle.
The system according to claim 2, wherein:

The license plate recognition device, the first sensor, and the roadside device are arranged at a target location that has the characteristics of slow driving of the vehicle and has a non-inductive charging requirement.
A trigger detection method, characterized in that it comprises:

Receive the road surface detection result of the road surface detection device and the roadside detection result of the roadside detection device, wherein the road surface detection device is used to detect the detection of the vehicle passing the target detection area through the first sensor buried in the target detection area on the road surface Event detection, and the roadside detection device is configured to detect an event when a vehicle passes through the target detection area through a second sensor installed on a roadside device that can cover the target detection area;

By fusing the road surface detection result with the roadside detection result, judging whether there is a vehicle passing by the target detection area;

Generate a trigger signal according to the judgment result.
The method according to claim 4, wherein:

The roadside detection result also includes location information of the detected vehicle;

The determining whether a vehicle passes by the target detection area includes:

If the road surface detection result indicates that a vehicle passes through the target detection area, and the roadside detection result indicates that the position information of the vehicle is within the target detection area, then the vehicle and the roadside included in the road surface detection result The vehicle in the detection result is determined to be the same vehicle, and it is determined that the vehicle is passing through the target detection area.
The method according to claim 4, wherein:

The road surface detection result includes: a first probability of a vehicle passing through the target detection area;

The roadside detection result includes: a second probability of a vehicle passing through the target detection area;

The determining whether a vehicle passes by the target detection area includes:

According to the first probability, the second probability and the judgment threshold information in the preset estimation algorithm, it is judged whether there is a vehicle passing by the target detection area.
The method according to claim 6, wherein:

The first probability is determined by the road surface detection device according to the initial detection result and the posterior probability corresponding to the road surface detection device;

The second probability is determined by the roadside detection device according to the initial detection result and the posterior probability corresponding to the roadside detection device;

Wherein, the posterior probability is determined according to the degree of matching between the initial detection result and the template signal and/or the trained mathematical model;

The initial detection result is: the respective detection system's perception result of whether there is a vehicle passing by the target detection area.
The method according to claim 7, wherein:

The judging whether there is a vehicle passing by the target detection area according to the first probability, the second probability and the judgment threshold information in the preset estimation algorithm includes:

Determine the first a priori probability corresponding to the road surface detection device and the second a priori probability corresponding to the roadside detection device according to the current weather and/or vehicle speed factors and the pre-stored historical statistical results corresponding to the respective detection systems The historical statistical results include: the confidence of the initial detection results of the respective detection systems under corresponding weather and/or vehicle speed conditions;

Add the product of the first probability and the first prior probability, and the product of the second probability and the second prior probability, and compare the result of the addition with the preset threshold information, It is determined whether a vehicle passes by the target detection area.
A trigger detection system is characterized by comprising:

A visual detection device is used to detect the event that a vehicle passes through the target detection area through a visual sensor installed on the roadside equipment and capable of covering the target detection area;

A radar detection device, which is used to detect an event that a vehicle passes through the target detection area through a radar sensor installed on the roadside equipment and capable of covering the target detection area;

The fusion algorithm processing device is used for fusing the visual detection result of the vision detection device with the radar detection result of the radar detection device, judging whether there is a vehicle passing by the target detection area, and generating a trigger signal according to the judgment result.
The system according to claim 9, wherein:

The fusion algorithm processing device is also used to send the trigger signal to a license plate recognition device, and the license plate recognition device is used to recognize the license plate of the vehicle.
The system of claim 10, wherein:

The license plate recognition device and the roadside device are arranged at a target location that has the characteristics of fast driving of the vehicle and has a demand for non-inductive charging.
The system according to claim 9, wherein:

The vision sensor is used to be installed on the roadside equipment by means of hoisting;

The radar sensor is used to be installed on the roadside equipment by means of hanging or side mounting.
A trigger detection method, characterized in that it comprises:

Receive the visual detection result of the visual detection device and the radar detection result of the radar detection device, wherein the visual detection device is used to detect the vehicle passing the target through the visual sensor installed on the roadside equipment and covering the target detection area Detection of regional events, and the radar detection device is used to detect the event of vehicles passing through the target detection area through a radar sensor installed on roadside equipment and capable of covering the target detection area;

By fusing the visual detection result with the radar detection result, judging whether there is a vehicle passing by the target detection area;

Generate a trigger signal according to the judgment result.
A trigger detection device is characterized by comprising:

The result receiving unit is used to receive the road surface detection result of the road surface detection device and the roadside detection result of the roadside detection device, wherein the road surface detection device is used to pass the first sensor buried in the target detection area on the road surface to detect vehicles passing by The event of the target detection area is detected, and the roadside detection device is used to detect the event of the vehicle passing the target detection area through a second sensor installed on the roadside equipment and capable of covering the target detection area ；

A fusion calculation unit, configured to determine whether a vehicle passes by the target detection area by fusing the road surface detection result with the roadside detection result;

The signal generating unit is used to generate a trigger signal according to the judgment result.
A trigger detection device is characterized by comprising:

The result receiving unit is used to receive the visual detection result of the visual detection device and the radar detection result of the radar detection device, wherein the visual detection device is used to pass the visual sensor installed on the roadside equipment that can cover the target detection area. The vehicle passes through the target detection area to detect the event, and the radar detection device is used to detect the vehicle passes through the target detection area through the radar sensor installed on the roadside equipment and can cover the target detection area ；

The second fusion calculation unit is configured to determine whether a vehicle passes by the target detection area by fusing the visual detection result with the radar detection result;

The second signal generating unit is used to generate a trigger signal according to the judgment result.
A roadside device, characterized in that it comprises:

The roadside detection device is used to detect the event that a vehicle passes through the target detection area through a second sensor installed on the roadside equipment and capable of covering the target detection area;

The fusion algorithm processing device is used for fusing the road surface detection result of the road surface detection device with the roadside detection result of the roadside detection device, judging whether there is a vehicle passing by the target detection area, and generating a trigger signal according to the judgment result; wherein The road surface detection device is used to detect the event that the vehicle passes through the target detection area through the first sensor buried in the target detection area on the road surface.
A roadside device, characterized in that it comprises:

A visual detection system is used for a visual sensor installed on a roadside device that can cover the target detection area, and detects events when a vehicle passes through the target detection area;

A radar detection system, which is used to detect an event that a vehicle passes through the target detection area through a radar sensor installed on the roadside equipment and capable of covering the target detection area;

The fusion algorithm processing device is used for fusing the visual detection result of the visual detection system with the radar detection result of the radar detection system, judging whether there is a vehicle passing by the target detection area, and generating a trigger signal according to the judgment result.