KR20140027661A

KR20140027661A - Ticketless toll gate system using plural camera

Info

Publication number: KR20140027661A
Application number: KR1020120093480A
Authority: KR
Inventors: 장태화
Original assignee: (주)메티스커뮤니케이션
Priority date: 2012-08-27
Filing date: 2012-08-27
Publication date: 2014-03-07

Abstract

A toll-free gate system using front and rear cameras is disclosed. The no-pass toll gate system includes a front image acquisition unit for acquiring a front image of the vehicle, a rear image acquisition unit for acquiring a rear image of the vehicle, and a laser sensor unit for acquiring height and length information of the vehicle. The apparatus may further include a vehicle number extracting unit configured to extract the number of the vehicle using the acquired front image and the rear image. The apparatus may further include a vehicle model information extracting unit extracting type information of the vehicle from the obtained front image, the rear image, and the height and length information of the vehicle. The vehicle charging unit may further include a vehicle charging unit configured to charge the toll for the vehicle using the extracted vehicle number and vehicle model information.

Description

Pass-free toll gate system using front and rear side cameras {TICKETLESS TOLL GATE SYSTEM USING PLURAL CAMERA}

The present invention relates to a toll gate system, and more particularly, to a toll gate system for setting a fare at a final destination toll gate after passing through the toll gate without the need for a ticket.

The highway toll gate system currently operated by Korea Expressway Corporation has a high-pass-based method and a general pass. The high pass-based system classifies vehicles by measuring the length and height of vehicles entering the high-pass lane. Currently, the high-pass lane can pass only passenger cars, vans and trucks of 4.5 tons or less.

In the general ticket issuing method, when a vehicle enters a general lane, the vehicle is recognized by recognizing the type of the vehicle and the license plate of the vehicle. To this end, the Korea Highway Corporation currently operates six types of vehicles, and the classification criteria are as follows.

Type 1: Two-axle vehicle. Width less than 279.4mm

Type 2: Two-axle vehicles, wheel width 279.4 mm or more, wheel width 1,800 mm or less

Type 3: Two-Axis Vehicle, Wheel Width Above 279.4mm, Wheelset Over 1,800mm

4 types: 3 axle truck

Five kinds: Special trucks with four or more axles

Class 6: Passenger cars, vans, cargo and special vehicles with a displacement of less than 1,000CC and a length of 3.6m and a width of 1.6m and a height of 2.0m and below

As such, the Korea Highway Corporation classifies the vehicle into six types by using data such as the number of wheels, the wheel width, the leap, and the height of the vehicle, and has a system for issuing and calculating a pass for the vehicle.

However, in order to classify vehicles based on the Korea Expressway Corporation's fare settlement standards, it is necessary to measure the number of shafts, wheel width, wheel width, wheel height (vehicle height), and so on. This method uses OVD (Object Vehicle Detector).

1 is a view schematically showing a state of use of the general toll gate system of Korea Expressway Corporation. As can be seen in FIG. 1, when the entry vehicle 108 enters the toll gate, it passes through the trapdle 100, which is a buried sensor, and measures the width, wheel width, and the number of axes, and further measures the height of the vehicle. The OVD 101 is operated to measure the yaw height. Based on the collected data, the vehicle class classification server 104 classifies the type of vehicle and instructs the pass issue machine 107 to issue a pass for the vehicle.

At this time, the evacuation prevention camera 103 takes a picture of the vehicle and the license plate in preparation for the vehicle type recognition error and transmits to the evacuation prevention server 105. The loading control camera 102 photographs a vehicle unconditionally and transmits it to the loading control server 106 when the entering vehicle is a cargo vehicle.

Since the existing toll gate system uses a buried sensor, construction has to be added when a new vehicle is released. For example, the Morning (1000cc) passenger car, which was not previously classified as a light car, was included as a light car when the related laws were amended. Figure 2 is a view showing the external size of the 1000cc light vehicle morning car.

However, even if the toll gate system is not changed even after the related laws are amended, Morningstar is not classified as a light car because it is larger than the previous light car, and thus a type 1 pass is issued. In order to prevent this, in order to recognize a passenger car such as a morning vehicle having a different size from the previous light vehicle, a separate trapdle must be installed.

However, since there are currently 1,800 lanes nationwide in Korea, KOGAS requires additional costs of about 54 billion won to be added if the additional cost of 30 million won per lane is needed to change the system. Will occur.

In addition, the standard for judging light cars is less than 1,000cc of displacement, and it is 3.6m in length and 1.6m in width and 2.0m in height and below. In principle, such a passenger car should be judged as one, but there is no method in the current system.

3 is a diagram illustrating an imported vehicle model in which a recognition error may occur. The vehicle illustrated in FIG. 3 has a displacement of 1,200 cc, thus deviating from the light vehicle standard, but the size of the vehicle meets the light vehicle standard.

As such, the conventional toll gate system has a problem in that the embedded sensor must be modified every time a new type of vehicle is released.

The present invention has been made to solve the above-mentioned conventional problems, and an object of the present invention is to provide a no-pass toll gate system that can be used even without a significant change in the system even when a new type of vehicle is released.

In order to achieve the above object, the pass-free toll gate system according to the present invention includes a front image acquisition unit for acquiring a front image of the vehicle, a rear image acquisition unit for acquiring a rear image of the vehicle, and a height and length information of the vehicle. It includes a laser sensor unit.

The apparatus may further include a vehicle number extracting unit configured to extract the number of the vehicle using the acquired front image and the rear image.

The apparatus may further include a vehicle model information extracting unit extracting type information of the vehicle from the obtained front image, the rear image, and the height and length information of the vehicle.

The vehicle charging unit may further include a vehicle charging unit configured to charge the toll for the vehicle using the extracted vehicle number and vehicle model information.

According to the present invention, since the pass is not issued, the time for stopping the vehicle to receive the pass at the highway entrance toll gate is greatly improved, thereby preventing the congestion of the highway. Improving the congestion of these entrance zones can greatly improve the environmental problems in the country, and greatly reduce the loss of time and cost due to congestion.

In addition, since the embedded sensor is not used, the cost of embedding the embedded sensor can be greatly reduced, thereby reducing the overall installation and operation cost.

What's more, when a new vehicle is driven, only the simple video data needs to be added, which simplifies the highway toll collection system.

If this system is used in conjunction with the mobile phone fare settlement system, it can be evolved into a high toll gate, which can be evolved into a high toll gate because it can be settled after payment without passing the fare at the exit toll gate. Can be.

1 is a view of a general toll gate system of Korea Expressway Corporation.
2 is a view of a 1000cc class light vehicle morning model.
3 is a view of an imported vehicle model in which a recognition error occurs.
4 is a diagram of a no-pass toll gate system structure using front and rear cameras.
5 is a detailed view of a no-pass toll gate system.
Fig. 6 is a detailed flowchart of operation of the pass-free toll gate system.
7 is a vehicle number recognition flowchart.
8 is a vehicle type recognition flowchart.
Fig. 9 is a view showing the operation of a toll-free gate.

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

4 is a view schematically showing a state of use of an embodiment of a toll gate system according to the present invention. Figure 4 shows the structure of the toll gate system with no pass using the front and rear cameras.

The present invention proposes a method using two cameras and one laser sensor as shown in FIG. 4 to solve the problem of the conventional toll gate and to recognize the type and number of the vehicle by using the image camera and the auxiliary sensor.

According to this method, the front camera 201 and the rear camera 200 can accurately recognize the type of the vehicle and the number of the vehicle based on the information obtained by simultaneously photographing the vehicle. By using vehicle model data and vehicle number information obtained from two cameras, the overall recognition rate is improved.

When the recognition rate of the first camera is 95% and the recognition rate of the second camera is 95%, the probability that each camera fails to recognize is 5%, so the probability that both cameras fail to recognize it is 5%. Is 0.025%, improving the overall recognition rate to 99.9975%.

Currently, Korea Expressway Corporation uses only the front camera to recognize the license plate of the vehicle, and shows an average recognition rate of 95%. The reason for the failure of recognition is caused by deliberate damage to the license plate, obstruction of the license plate by snow or dust, or distortion of the license plate.

In order to damage the license plate deliberately, numbers may be manipulated or special paint may be applied. Sometimes, the license plate may be driven in a dusty condition without cleaning the license plate.

In fact, it is very rare for a vehicle that intentionally damages a license plate in the field, so it is very rare to damage both the front license plate and the rear license plate. If the front license plate is not recognized, the rear license plate can be recognized to improve the overall number recognition rate. have.

To this end, in the present invention, two video cameras are used to realize a pass-free toll gate system that does not require a separate reclamation work. In order to improve the vehicle number recognition and the vehicle model recognition rate, the front camera 201 and the rear camera 200 may be used to obtain information about the entered vehicle as much as possible.

In the present invention, there is a technical focus on photographing an entry vehicle at the same time by installing two cameras on the front and rear sides for the vehicle number recognition and the vehicle type, which are the most important factors in realizing the toll gate system. .

In order to maximize the recognition probability when two cameras are simultaneously photographing the entry vehicle, three shots are continuously performed, and the number of vehicles and the main feature points of the vehicle are extracted from three consecutive shots. The extracted number and feature point data are automatically transmitted to the server. At this time, the camera located on the rear side is installed to shoot between the angle of 15 to 45 degrees to obtain the optimum shooting angle.

FIG. 5 is a view illustrating in more detail the state of use of the toll-free gate system of FIG. 4. In FIG. 5, it can be seen that the no-pass toll gate system includes a front camera 201, a rear camera 200, and a laser sensor 202.

6 is a detailed flowchart of the toll gate system. In FIG. 6, first, when the vehicle enters, the front camera and the rear side camera photograph the entered vehicle at the same time. Subsequently, the license plate area is extracted from the images of the vehicle photographed by the front camera and the rear camera.

When the license plate area is extracted, the vehicle number is extracted to recognize the correct number in the license plate area. 7 is a flowchart of a vehicle number recognition process. The process of extracting the vehicle number is similar to the existing license plate recognition method.

If it fails to recognize the license plate of the vehicle by using the front camera, the number of the vehicle is recognized by using the rear license plate recognized by the rear camera.

After the vehicle number is recognized as described above, the type of vehicle is again classified. The classification of the vehicle type is to extract the main feature points of the vehicle and compare it with the standard image of the vehicle.

8 is a flowchart of a vehicle type recognition process. In FIG. 8, a vehicle classification method using the feature points of the vehicle is represented. From the captured images, feature points for radiator grille, headlights, taillights and rear trunk parts, which are the main points that characterize the vehicle, are extracted in advance.

To this end, feature information about the front and rear sides is extracted from image data of an existing vehicle and stored in a database. When the vehicle enters the toll gate, the front camera and the rear camera extract only data corresponding to the main feature points from the captured images. If this data is matched more than 90% compared with the previously stored standard data, classify the vehicle.

For example, a standard image of the 2002 Grandeur was taken on the front, side, and rear, and the main features of the vehicle were stored in the database. If the 2002 Grandeur entered the toll gate, the front, rear, and side views were taken. It extracts data from the part and compares them to match the car model.

Since only the feature point information is extracted from the main part of the vehicle, the matching process for classifying the vehicle is shortened, and it provides an advantage that highly accurate vehicle classification is possible.

Through the above process, the number of the vehicle and the type of the vehicle are measured from the image information of the entrance vehicle obtained from the front camera and the rear camera, and the vehicle height information is input from the laser sensor as an auxiliary means. The car models are classified according to the 6 types of vehicle classification criteria.

If the numbers extracted from the front and rear cameras match the vehicle model, they are sent to the server normally. If there is a difference in the extracted data, the data is transferred to the server for visual reading.

9 is a view for explaining the operation of the pass-free toll gate system according to the present invention. The system of Figure 9 is a system that can pass through the highway without a pass without escaping from the existing toll gate method for issuing and paying the right of passage.

First, the step of acquiring the information of the vehicle proceeds to the next step.

In the first step, when the vehicle enters, the front camera 201 extracts feature points such as a license plate and a radiator grille and headlight of the vehicle from the front of the vehicle.

In the second step, the rear camera 200 extracts feature points such as the rear license plate, the rear bumper, the tail light, and the number of axes of the vehicle from the rear side of the vehicle.

In step 3, the height of the vehicle is measured by the laser sensor 202.

In step 4, each piece of information is transmitted to the integrated management server 203.

In step 5, the integrated management server 203 extracts the vehicle number and model. The vehicle number is extracted from the license plate area of the vehicle using the existing image recognition algorithm. The classification of the vehicle type is based on the standard image data of the previously stored vehicle and the characteristic data of the currently entering vehicle. Use a method of classifying cars by comparing them. As of 2011, about 400 kinds of vehicles are operating in Korea, and standard image information of these vehicles can be extracted by using image data of vehicles secured by the Korea Expressway Corporation.

In step 6, the extracted license plate and vehicle model data are stored in the integrated management server 203.

In step 7, if the entering vehicle enters the exit toll gate again to exit the highway after driving the destination, the process of steps 1 to 6 is repeated. The integrated management server 203 determines where the current vehicle enters and exits by using the vehicle model and number data raised from the exit direction toll gate, and determines and informs the corresponding fare.

Step 8 ends with the settlement of the final fee.

The present invention relates to a no-pass toll gate system that automatically recognizes the number of a vehicle and the type of the vehicle using two cameras, passes through the toll gate without the need for a ticket, and sets the fare at the final toll gate.

The highway toll gate system operated in Korea used a separate sensor to measure the height and length of the vehicle to distinguish the type of vehicle when the vehicle enters the toll gate.

The length of the vehicle is measured by a machine called a trapdle using a loop sensor embedded in the road itself, and the height of the vehicle is measured by an object vehicle detector (OVD) laser sensor.

Existing toll gates issue passes with the type of vehicle using the trapdle and the OVD. Such a method has a problem in that a lot of costs in construction because it uses a buried sensor embedded in the road.

In addition, when a new vehicle is released, a problem arises in that construction must be performed again accordingly. For example, in the past, only a 800cc passenger car was recognized as a light car, but as a 1,000cc passenger car is recognized as a light car, a vehicle such as a 1,000cc morning car (Kia Motors) will enter the range of light vehicles. In such a case, the existing system must proceed with the addition of a buried sensor to recognize 1,000cc morning.

On the other hand, the toll gate system using the two cameras and the height measuring sensor proposed in the present invention is a system for recognizing the type and number of the vehicle from the images obtained by photographing the vehicle entering from the front and rear sides of the vehicle.

There is no need to stop the vehicle as it enters the vehicle and recognize the vehicle's type and number through video signal analysis and pass this information to the Korea Expressway Corporation's wide area server. It is a system to settle.

Two cameras simultaneously acquire and acquire images from the front and rear sides to acquire key information about the vehicle, extract important features of the vehicle from the acquired images, and compare the images with those of previously stored vehicles. Classify.

In this case, in order to provide a more accurate recognition rate, a laser sensor for measuring a separate vehicle height is used. In this way, the performance of recognizing the type of vehicle can be improved to 99.9% or more.

In addition, by simultaneously photographing the license plate of the vehicle from the front and rear side, the recognition rate for the vehicle license plate can also be improved to 99.9% or more. As such, by accurately recognizing the type of the vehicle and the number of the vehicle, it is possible to implement a no-pass system that can pass through the toll gate without issuing a separate pass.

Apart from this, by simultaneously photographing the vehicle from the front side and the rear side, the loading failure of the freight vehicle can be automatically read. This function is able to automatically recognize whether there is a poor loading by comparing the acquired image of the cargo compartment of the entering vehicle with a standard image of the cargo vehicle.

The toll gate system using two cameras and height measurement sensor proposed in the present invention is a next generation toll gate system that can simultaneously classify vehicles, recognize license plates, and control loading failure using one system.

Although the present invention has been described in terms of some preferred embodiments, the scope of the present invention should not be limited thereby but should be modified and improved in accordance with the above-described embodiments.

100: trapdle
101: Object Vehicle Detector (OVD)
102: intermittent camera
103: evasion prevention camera
104: vehicle classification server
105: evasion prevention server
106: Load Enforcement Server
107: ticket machine
200: rear camera
201: front camera
202: laser sensor
203: Integrated Management Server

Claims

A front image acquisition unit obtaining a front image of the vehicle;
A rear image acquisition unit for acquiring a rear image of the vehicle; And
The toll gate system of claim 1, further comprising a laser sensor unit for acquiring vehicle height and length information.

The method of claim 1,
And a vehicle number extracting unit which extracts a number of the vehicle using the obtained front image and the rear image of the vehicle.

3. The method of claim 2,
And a vehicle model information extracting unit which extracts vehicle type information from the obtained front image, rear image, and height and length information of the vehicle.

The method of claim 3,
The toll gate system of claim 1, further comprising a vehicle charging unit configured to perform toll charging for the vehicle using the extracted vehicle number and vehicle model information.