JPS5870166A - Method for measuring traveling time of vehicle - Google Patents

Abstract

PURPOSE:To measure a time taken in traveling of a vehicle, by arranging image sensors such as television cameras in plural observation positions and comparing and collating picture information of sampled vehicle numbers of vehicles passing through observation positions to obtain the matching of patterns of vehicle numbers. CONSTITUTION:An image sensor ISA is arranged 5m above the ground in the about center of one traffic lane at a position A. When the arrival of a vehicle C at a detecting line DL is detected on the picture, only a partial picture including a number plate NP is cut on the basis of detection information of the vehicle and is stored in a picture memory. A binary picture of the number plate is obtaind, and its binary luminance is processed with a slice level of the luminance in a processing device A3, and the picture subjected to the binary encoding processing is transmitted to a center M through a high-speed transmission line. Through the similar processing course, binary information of the number plate is transmitted from a position B to the center M for every arrival of a vehicle. In the center M, number plates are collated on the basis of picture information transmitted from positions A and B, and the time when the vehicle passing through the position A passes through the position B, and thus, a time taken in traveling of the vehicle C from the position A to the position B is measured.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is characterized by disposing image sensors such as television cameras at a plurality of observation points and comparing and collating image information of sampled car numbers of vehicles passing through the observation points. It relates to a method for measuring vehicle travel time in road traffic, and is intended to contribute to improving the control effectiveness of traffic control systems.

Traffic control centers located mainly in major cities across the country perform optimal control of vehicle traffic flow in 411 aspects, thereby reducing vehicle travel time, achieving macro-economic effects, and reducing pollution such as exhaust gas. , has been highly effective in reducing traffic accidents. Although this traffic control system has various control measures using one software, there is no effective means to measure vehicle travel time, which is an evaluation index of control effectiveness, so it is difficult to find the optimal control calculation. The current situation is that it is difficult to make an optimal evaluation based on the law.

Now, to explain the second part of the conventional method of measuring the traveling time of a vehicle, as shown in Fig. 1, a person is placed at two points A and B on the route of one vehicle, and the vehicle C is focused on. car number,
A manual method of measuring the traveling time of vehicle C traveling distance l between points A and Ba by communicating the vehicle type, color, etc. using the portable radio AI'*E1, or a method of measuring the travel time of vehicle C traveling distance l between points A and Ba, or by communicating both the vehicle and the road. In the case of having a wireless communication channel for both directions, as shown in Fig. 2, roadside radio equipment A is installed at two points A and B on the road.
2. B2 is installed, and a specific code co is sent from the vehicle C to the traffic control center M via the roadside machine A2 t B2.
There is a method of using a route guidance system that calculates the traveling time of vehicle C traveling distance l between two points A and B.

However, the manual method described above cannot be used for real-time measurement of traffic control systems, and the method using the route guidance system shown in Figure 2 is not suitable for practical use because it requires installing in-vehicle radio equipment in many vehicles. Each has its own drawbacks, such as being short-lived.

The present invention provides a method for measuring vehicle travel time in a traffic control system by using an image sensor that is in the realm of practical use, such as a television camera. Image sensors I8A and ISB are installed at two points A and E, respectively, and the processing device A3 uses the vehicle numbers of vehicles arriving at each point as image information.
Alternatively, the images are sent to the traffic control center (hereinafter simply referred to as the center) M via B3, and the center M compares and collates the image information at points A and B2, so that the vehicle C of interest is A,
The time taken to travel the distance 1lIl between two points B is calculated.

Hereinafter, an embodiment of the traveling time measuring method according to the present invention will be explained with reference to the drawings. Figure 4 shows a single lane running from the left to the right on the traveling road shown in Figure 3, where the width of each lane is 3.5 mm. 2 is a perspective view showing a soft part in which an image sensor ISA is disposed at a position 5 m above the approximate center of the lane at point A, and its vehicle detection line DL is set on the lane. FIG. If a television camera is used as the image sensor ISA, image information obtained by the television camera of the vehicle C that has arrived at the vehicle detection line DL is shown as shown in FIG. In the case of a scanning television camera, such an image takes a time of 1/30 sec nil tI rft (0
Obtained as a percentage. When it is detected on the image that the vehicle C has arrived at the detection line DL, one screen of image information at that time is stored in the image memory.

However, in the case of this method, it is not necessary to memorize the entire screen. That is, the image information of the portion surrounded by the detection line DL and the line (a) at the height from DL and the line (b) and (c) on both sides that include the number play) NP shown in FIG. -D (analog-digital) conversion and storage is sufficient. The license plate NP attached to the front of vehicle C is 30 = 40 cm + in the case of a regular car.
.

In the case of large vehicles, it is approximately constant in the range of 60 to 70 CI11. Furthermore, the attachment point to the vehicle can be considered to be within ±l cm from the center of the vehicle to either the left or right. Furthermore, the mounting inclination angle is also in the range of ±10 to 20°.

For the above reasons, based on the vehicle detection information, only the partial image containing the number play (NP) is cut out from the image shown in FIG. 5 and stored. Therefore, the fourth
Assume that one television camera is installed to monitor the 1st lane, like the image sensor I8A in the figure. As shown in Figure 4, the lane width is 5.5m, and one screen is 512 x 51
When configured with a pixel resolution of 2, the resolution is 070m71 pixels. Note that license plates do not have an infinite number of sizes, but are generally classified into several types. FIG. 6 shows an example of the size of the license plate NP. As shown in the figure, the thickness of the characters is 12wI, so in order to increase the resolution by one level, it is necessary to zoom in on the image in FIG. 5 as shown in FIG. 7. Therefore, if we zoom in on the image in Figure 5 with a lane width of 3.511 and enlarge it to an image representing the surface of LIIIXIIll, and configure this with a resolution of 512 x 512 pixels, it will be approximately 0.2 cm x 0.2 cm71
It has a pixel resolution, which is sufficient.

Next, out of the enlarged image above, number play) N
Extract only the P part as shown below. i.e. 1
The white part of the number low rate NP has high brightness, so this part is set to the logic value "1", and the black text part is set to -HI.
If I+, a binary image of the license plate is obtained as shown in FIG. This binarized luminance is processed at the luminance slice level in the processing device A3 shown in Fig. 3, and the binary encoded image is transmitted to the center M through a high-speed transmission line ■Similar processing After the process, B
Binarized license plate information is sent to the center M for each arriving vehicle from the location. Center M verifies the license plate based on the image information sent from both points A and B, learns the time when the vehicle that passed point A will pass point B, and identifies the time when the vehicle that passed point A will pass point B. It measures the travel time between the two points.

In the above case, the method of the present invention does not involve the complicated process of decoding the letters and numbers on the license plate.
As will be explained with reference to FIGS. 9 and 10, the size of license plates is standardized to directly match patterns. Standardizing the size means that license plates generally have a ratio of height: width - 1 = 2.
It is overwhelmingly 0-80%, so it is vertical.

The horizontal parts are subdivided into meshes at the same ratio, so that the total sum of the subdivisions is equal.

Next, referring to FIG. 9, (5) in the same figure.

■) are both images of the same license plate, and (4) in the same figure
Let be the photographed image NPa at point A, and let CB in the figure be the photographed image NPb at point B. By the way, when a progressive scanning TV camera is used for the image sensors ISA and ISB in Fig. 3, the scanning time for one screen is approximately 1/30 sec, but the scanning time for the license plate portion is 115 seconds, or approximately 1/30 sec. Completes in 150 seconds. Assuming that the traveling speed of the vehicle is 6 o Mh, the vehicle will be approximately I during the scanning time.
It will move Qcm. However, if a camera with a height of 51++ is used to photograph the vehicle at a distance of 10 to 15 fN in front of the vehicle, an image suitable for pattern matching can be obtained. Therefore, the image areas of FIG. 9(4), Φ) are not necessarily equal and must be different. However, it goes without saying that the two are similar.

Therefore, each screen NP21. When NPb is subdivided into areas proportional to the similar areas, the sum of each subdivision number is equal, so the intersection of the meshes in Figure (4) ``1pa2 t a5''
* "" and the mesh (7) of the same figure intersect b1. b2゜bs+... correspond to each other, and their sums are equal to each other. Therefore, mutually corresponding points a1 and bl, a2 and b2 . a3 and b5゜...
Then, the binarized logical values of brightness are multiplied with each other to obtain their sum DO=Σaibi for one entire screen. FIG. 10 is a flowchart showing the processing process. The product aibi of the logical values of corresponding points from the standardized images NPa and NPb is calculated, and if the sum DO is above the threshold level, the pattern of the images NPa and NPb is determined. indicates that a match is confirmed.

As explained in the above embodiments, according to the method of the present invention, vehicle running time can be measured by finding the matching of car number patterns without using advanced technology for decoding car numbers. It is possible.

In addition, when point B has multiple lanes, the image sensor is
Several units may be installed and the video signals may be switched and sequentially processed. In general, there are many causes of deterioration of S/N such as dirt, distortion, and blurring of the license plate, so it is difficult to decipher the car number itself. A method based on pattern matching eliminates such difficulties. Furthermore, when measuring the travel time, it is not necessary to apply it to all arriving vehicles, but only to a sampling of vehicles, so if you select a vehicle number that can be clearly recognized, the effectiveness of this method will be further enhanced. be done.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram showing an example of a method for measuring vehicle travel time manually; FIG. 2 is an explanatory diagram showing an example of a method for measuring vehicle travel time using a two-way wireless communication channel between the vehicle and the road; Third
The following figures show an embodiment of the vehicle running time measuring method of the present invention. Fig. 3 is an explanatory diagram showing an overview of the equipment arrangement state, and Fig. 3 shows the arrangement of the image sensor for setting the vehicle detection line. A schematic diagram, FIG. 5 is a schematic image diagram of a vehicle that has arrived at the detection line,
Figure 6 is an example of the size of a license plate, Figure 7 is an example of the size of a license plate.
The figure is a zoomed-up view of the vehicle image in Figure 5, Figure 8 is an explanatory diagram of binarization of license plate brightness, Figure 9 is an explanatory diagram of standardization of license plate size, and Figure 10 is a license plate pattern. It is a flowchart showing a confirmation processing process. L: Vehicle driving route, A, B: Required point, I8A. ISB: Image sensor, DL: Vehicle detection line,
C: Vehicle, NPC license plate, NPa, NP
b: License plate pattern. Patent Applicant: Nippon Signal Co., Ltd. Attorney: Rikichi Kawa, Figure 1, Figure 2I21, Figure 3,

Claims (1)

[Claims] A vehicle detection line is set by disposing an image sensor at a plurality of required positions on a vehicle travel route, and a license plate image of a vehicle is extracted from an image taken by the image sensor of a vehicle arriving at the detection line. The brightness of the extracted image is then converted to a binary logical value and standardized to enable matching, and the standardized vehicle license plate patterns obtained from the plurality of points are compared. A method for measuring vehicle travel time, characterized in that it is possible to measure vehicle travel and traveling time.