JPH04141800A - Vehicle kind decision device - Google Patents

Abstract

PURPOSE:To prevent the misreading of the kind of a vehicle to the one with a toll higher than the original one by performing higher weighting in order of the kind of the vehicle with lower toll against the probability of the decision result corresponding to each kind of the vehicle at the time of the vehicle kind decision for vehicle. CONSTITUTION:When a vehicle kind decision part 6 decides the kind of the vehicle by utilizing information such as the vehicle kind code of a classification number, plate size, vehicle length obtained from a vehicle measuring device 5, the number of shafts, the number of rings, or the like, the decision result is transferred through an automatic ticket machine interface 7 to an automatic ticket machine 8, and the toll ticket for the toll corresponding to the kind of the vehicle is issued. When the vehicle kind decision for vehicle is performed, the higher weighing in order of the vehicle kind with lower toll is performed against the probability of the decision result corresponding to the respective kinds of vehicles. Thus, the worst case of misreading the kind of the vehicle to the one with higher toll can be prevented, and a user does not feel discomfort and displeasure.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention provides a method for calculating a toll fee corresponding to the type of vehicle passing through, for example, at an entrance gate of a toll road such as an expressway. The present invention relates to a vehicle type discrimination device that automatically discriminates the type of vehicle.

(Prior Art) Conventionally, for example, at the entrance gate of a toll road such as an expressway, in order to issue a toll ticket (calculate the toll) corresponding to the vehicle type of the passing vehicle, the vehicle type of the vehicle is automatically determined. There is a vehicle type identification device. For example, as disclosed in Japanese Patent Application Laid-open No. 61-213998, there is a well-known technology for automatically identifying vehicle types using information such as the classification number present on the vehicle's license plate and the size of the license plate. be.

The classification number is an Arabic decimal number of two digits or less that is present on the license plate in the format shown in FIG. 9, and has the meaning as described in the Road Transport Vehicle Law Enforcement Regulations. Here, Figure 9 shows the format of license plates attached to vehicles such as automobiles in Japan, where a is a serial vehicle number consisting of Arabic large numbers of 4 digits or less, and b is a serial vehicle number of 2 digits or less. The classification number consists of Arabic decimal digits, C is a kanji code indicating a land transportation branch, and d is a kana/roman code indicating whether or not it is used for transportation business. Note that, of the classification number, the left-digit number that is particularly effective for identifying the vehicle type is called a vehicle type code e for convenience.

When reading the classification number present on such a license plate in a reading process, errors may occur in the reading result. A particular problem is when a car type is misread as having a higher toll than the original car type.

(Problems to be Solved by the Invention) As described above, in the conventional vehicle type discrimination process, misreading may occur, for example, in the process of reading the classification number on the license plate. If the frequency of such misreading is within a certain range, it will not be a big problem for the operator, or it will not be a big problem for the user. For users, is it still a good thing if a ticket is issued for a vehicle type with a lower toll than the original vehicle type?If a ticket is issued for a vehicle type with a higher toll than the original vehicle type, the frequency of such tickets may be worse. Even a small amount is a problem.

Therefore, the present invention avoids the worst case of misreading the information indicating the vehicle type such as the classification number on the license plate, in which the vehicle type is mistaken for a high toll vehicle type. The purpose of the present invention is to provide a vehicle type discrimination device that does not cause psychological discomfort or dissatisfaction to users.

[Structure of the Invention] (Means for Solving the Problems) The vehicle type discrimination device of the present invention can be used on toll roads.
In order to calculate the toll fee corresponding to the type of vehicle passing through,
A vehicle type discrimination device that automatically discriminates the vehicle type of the vehicle, comprising an imaging means for capturing a scene image including a license plate of the vehicle, and a detection unit for detecting a license plate area from image information captured by the imaging means. a recognition means for detecting and recognizing information representing the vehicle type from the image information of the license plate area detected by the detection means; and a recognition means for detecting and recognizing information representing the vehicle type from the image information of the license plate area detected by the detection means; The vehicle includes a weighting means for weighting, and a determining means for determining the type of vehicle using the recognition result weighted by the weighting means.

(Function) According to the vehicle type discriminating device of the present invention, when discriminating the vehicle type of a vehicle, the probability of the discrimination result corresponding to each vehicle type is weighted in descending order of the vehicle type with the lowest toll. Even if a misreading occurs in the process of reading information representing the vehicle type, such as a classification number, it is possible to easily identify the vehicle type with a higher toll.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 1 schematically shows the configuration of a vehicle type discrimination device according to the present invention. That is, the image capturing unit 1 captures a forward image including the license plate of an incoming vehicle at an entrance gate of a toll road such as an expressway. The license plate information detecting unit 2 analyzes the image information captured by the image capturing unit 1, and detects a classification number relative to the serial vehicle number existing on the license plate. The classification number recognition unit 3 is the license plate information detection unit 2.
The classification number detected by the method is recognized.

The plate size determination unit 4 determines the size of the license plate. The vehicle meter 41 device 5 measures the length, number of axles, number of wheels, etc. of the vehicle using, for example, a pressure sensor or a photo sensor. The vehicle type discrimination section 6 discriminates the vehicle type based on the classification number obtained from the classification number recognition section 3, the plate size obtained from the plate size discrimination section 4, and the measurement information obtained from the vehicle measuring device 5.

The automatic ticket issuing machine interface 7 is an interface for transmitting the vehicle type information determined by the vehicle type determining section 6 to the automatic ticket issuing machine 8. The automatic ticket issuing machine 8 automatically issues a toll ticket with a toll corresponding to the vehicle type.

The memory unit 9 is a general-purpose storage means used for various processes, and the CPU 10 mainly controls the entire process. The general-purpose bus 11 is a data bus used for general-purpose data transfer, and the image bus 2 is a data bus used exclusively for transferring image information.

Furthermore, details of each component will be described. First, the image capturing section 1 is configured as shown in FIG. 2, for example. That is, a television camera (hereinafter simply referred to as a television camera) 21, a near-infrared flash device 2
2. Photo sensor light emitting section 23a1 Photo sensor light receiving section 2
3b, an A/D converter 24, and an image memory 25, and is installed at the entrance gate of a toll road. As shown in FIG. 3, the photosensor light receiving section 23b detects that the light emitted from the photosensor light emitting section 23a is blocked due to the vehicle 27 entering the entrance gate 26 of the toll road, and the detection signal is detected by the photosensor light receiving section 23b. A near-infrared flash device 22 emits near-infrared flash light, and a television camera 21 captures and captures a front image of the vehicle 27 including the license plate. This captured image information is converted into, for example, an 8-bit digital signal by the A/D converter 25,
The image is temporarily stored in the image memory 25.

The near-infrared flash light emitted by the near-infrared flash device 22 may be used to stabilize illumination conditions or to be used as night illumination, or if visible light flash light is emitted near the vehicle as it may disturb the driver. Infrared light is used.

For example, as shown in FIG. 4, the license plate information detection section 2 includes three binarization processing sections, a serial vehicle number detection section 32,
and a classification number detection section 33.

The binarization processing unit 31 converts the captured image stored in the image memory 25 into two formats, for example, a character is "1" and a background is "0".
Convert to value image. When there is a contrast between the characters and the background, such as in a license plate image, binarization processing using contrast detection is effective. As such a binarization method using contrast detection, for example, the technique disclosed in Japanese Patent Application Laid-Open No. 63-300744 can be used.

The serial vehicle number detection unit 32 detects, for example, a candidate area of the serial vehicle number that is easiest to find among license plate information. Detection of candidate areas for serial vehicle numbers is performed, for example, by applying labeling processing to the binary image, detecting connected areas connected by "1", and then determining the vertical and horizontal sizes of the detected connected areas, the arrangement of each connected area, etc. This is done based on the analysis results. The labeling process, for example, as shown in FIG.
A label is attached to each connected region, and the coordinates XSI to XS4 of a rectangle circumscribing each connected region are provided as accompanying information.
YSI ~YS4. XEI-XE4. YEI〜Y
E4 is required. Since the vehicle images are taken at approximately the same point by the photosensor, the size of the serial vehicle numbers on the images approximately falls within a certain range of variation. In addition, the arrangement is based on the size regulations for license plates stated in the Enforcement Regulations of the Road Transport Vehicle Act. Vehicle number candidates can be detected.

Further, the classification number can be detected relatively easily from the relative positional relationship based on the size regulations for license plates described in the Road Transport Vehicle Law Enforcement Regulations.

The classification number detection unit 33 performs classification based on the coordinate information of the serial vehicle number candidate area obtained from the serial vehicle number detection unit 32 based on the relative positional relationship using, for example, the same means as for detecting serial vehicle number candidates. Attempt to detect number. Note that detection of the classification number also serves as a means of verifying serial vehicle number candidates. That is, when the constituent elements of the serial vehicle number and classification number are true, the area is output as the true license plate area.

The classification number recognition unit 3 performs recognition processing of the detected classification number, for example, by the well-known rZ turn matching method. Using the coordinate information obtained from the license plate information detection section 2, a numerical pattern is cut out from the binary image, and the cutout pattern is compared with the reference tZ turn to determine the degree of similarity. The degree of similarity is a numerical measure that indicates how closely the input pattern and the reference pattern are related, and here, for convenience, it takes a value from 0 to 100, and the degree of similarity is a value between 0 and 100. It is assumed that

The plate size determining section 4 determines whether the size of the number plate is large or medium, using the technique disclosed in Japanese Patent Application Laid-Open No. 63-1.05916, for example.

The vehicle type discrimination unit 6 uses the vehicle type code and plate size of the classification number obtained using the above components, and information such as the vehicle length, number of axles, and number of wheels obtained from the vehicle measuring device 5. The vehicle type is determined using the flowchart shown in FIG. 7, for example. The determination result is transmitted to the automatic ticket issuing machine 8 via the automatic ticket issuing machine interface 7 Ace 7, and a toll ticket corresponding to the vehicle type is issued. By the way, tolls will be paid in order of light vehicles, motorcycles, ordinary, large, and extra large.For non-vehicles, a toll collector will collect the toll at the exit gate.

The vehicle type is automatically determined by the processing described above. However, in processes such as the above-mentioned automatic reading of vehicle model codes, reading errors may occur. In that case, a problem arises when the vehicle type is mistakenly read as having a higher toll than the original vehicle type. Therefore, in the present invention, the reliability of the reading result (recognition result) for each vehicle type code is weighted higher in the order of vehicle types with lower tolls.

For example, it is assumed that the degree of similarity for the reading of vehicle model codes "0 to 9" is obtained as shown in FIG. 7, for example. In the case of this example, there is almost no difference in the degree of similarity between the recognition results of "2" and "7". In terms of simple size relationship, the similarity of "2" is slightly high, and if it has two axles, it will be determined as medium-sized, or if it is actually "7", the price will be higher than the original car type. In such a case, for example, by weighting each degree of similarity as shown in FIG. 8, the degree of similarity of "7" becomes slightly higher, and the vehicle type is determined to be normal. Note that this weighting process may be performed, for example, by the classification number recognition unit 3 or by the vehicle type determination unit 6.

Therefore, even if the reading result is incorrect, the vehicle type will not be charged a higher toll than the original vehicle type. In this way, it is possible to identify the vehicle type without misreading the vehicle type as having a higher toll than the original vehicle type.

In the embodiments described above, the vehicle is detected by a photo sensor, but it does not matter whether the vehicle is detected by, for example, an ultrasonic sensor or by image processing. Further, the captured image information may be quantized to 8 bits, but the number of quantized bits is not limited to this.

Further, the binarization method is limited to the method disclosed in Japanese Patent Application Laid-Open No. 63-300744. For example, binarization using threshold processing may be used. Further, detection of license plate information is not limited to detection of connected regions using labeling processing, but may also be based on projection information on the X-axis and Y-axis, which is a general character string detection method.

Furthermore, the recognition of the license plate information (classification number) is not limited to the pattern matching method, but may also be recognized by structural analysis, which is a general character recognition means. However, weighting of the certainty of recognition results is not limited to similarity. Furthermore, it goes without saying that the weighting is not limited to the coefficients shown in FIG. Furthermore, the procedure for determining the vehicle type is not limited to that shown in the flowchart of FIG.

[Effects of the Invention] As detailed above, according to the present invention, when determining the vehicle type, the certainty of the classification result corresponding to each vehicle type is weighted in descending order of toll toll. By doing this, even if a misreading occurs during the reading process of information indicating the vehicle type such as the classification number, the worst-case scenario of mistaking the vehicle type for a higher toll can be avoided, and the system will provide psychological protection to the user. It is possible to provide a vehicle type discrimination device that does not cause discomfort or dissatisfaction.

[Brief explanation of the drawing]

1 to 8 show an embodiment of the present invention,
FIG. 1 is a block diagram showing the overall configuration, FIG. 2 is a block diagram showing the configuration of the image capturing section, FIG. 3 is a diagram explaining vehicle image capturing, and FIG. 4 is a block diagram of the license plate information detecting section. A block diagram showing the configuration, FIG. 5 is an explanatory diagram explaining the labeling process, FIG. 6 is a flowchart explaining the vehicle type discrimination process, FIG. 7 is a diagram showing an example of similarity corresponding to the vehicle type code, and FIG. 8 9 is a diagram showing weighting coefficients corresponding to each car type coat, and FIG. 9 is a diagram showing the format of a license plate attached to a car. 1... Image capturing unit, 2... License plate information detection unit, 3... Classification number recognition unit, 4...
...Plate size discrimination unit, 5...Vehicle measuring device, 6...Vehicle type discrimination unit, 7...Automatic ticket issuing machine interface, 8...Automatic ticket issuing machine, 9.・・Memory part, ○ ・・CPU. 1...General purpose bus, ri...image lotus.

Claims (1)

[Scope of Claims] A vehicle type discrimination device that automatically determines the vehicle type of a passing vehicle in order to calculate a toll fee corresponding to the vehicle type of the passing vehicle at a toll road gate, the device comprising: an imaging means for capturing a scene image including a scene image; a detection means for detecting a license plate area from the image information captured by the imaging means; and a detection means for detecting information representing a vehicle type from the image information of the license plate area detected by the detection means. a recognition means for recognizing the recognition result of the recognition means; a weighting means for weighting the certainty of the recognition result of the recognition means in ascending order of toll; and a recognition result weighted by the weighting means to identify the vehicle type of the vehicle. A vehicle type discriminating device comprising a discriminating means for discriminating.