KR20040005299A - Grouping apparatus for automobile and method thereof - Google Patents

Abstract

PURPOSE: An apparatus for discriminating the type of a vehicle and a method for the same are provided to discriminate the type of the vehicle without stopping the vehicle by only using the photographed image. CONSTITUTION: An apparatus for discriminating the type of a vehicle(1) includes a vehicle detector(2), a first camera(3), a second camera(4) and an image processor(5). The vehicle detector(2) detects the vehicle approach(1). The first camera(3) is installed on the top of the approaching vehicle(1) and the second camera(4) is installed on the side of the intruding vehicle(1). And, the image processor(5) extracts the contour of the top portion and the side portion from the image photographed from the first and the second cameras(3,4) and measures the height, length, width and errors therefor from the extracted contours.

Description

Vehicle discrimination apparatus and method {GROUPING APPARATUS FOR AUTOMOBILE AND METHOD THEREOF}

The present invention relates to a vehicle type determination apparatus and method, and more particularly, to a vehicle type determination apparatus and method for determining the type of the vehicle passing through the collector in the toll collection system to collect the fare corresponding to the type of the vehicle.

In general, a system for collecting a fee on a toll road may collect a certain fee regardless of a vehicle type or differentially collect a fee according to a vehicle type.

In a system that collects a certain amount of money regardless of the vehicle type, it is only necessary to confirm the payment of a fee such as a coin thrower, but in a system that needs to collect the differential fee according to the vehicle type, it is necessary to distinguish the vehicle type.

Currently, the Toll Collection System (TCS), which is commonly used for toll collection on highways, is a method in which a driver issues a magnetic card at an entry toll gate and pays a fee at an arrival toll gate.

Such a method deepens the traffic congestion at the toll gate and causes problems such as increased logistics costs and environmental pollution.

In order to solve the problem of the manual toll collection system, an electronic toll collection system (ETCS) has recently been proposed.

This electronic toll collection system (ETCS) is a computerization of toll collection and wirelessly using dedicated small region communication (DSRC) while the vehicle is traveling without stopping when the vehicle passes through the toll gate. It is to handle the collection of charges.

However, the method of wirelessly collecting tolls cannot distinguish between a vehicle that pays a toll and a vehicle that does not pay only by wireless communication, and attaches an OBU of a passenger car to a bus, such as an On Vehicle Unit (OBU) of another vehicle. You can't prevent traffic through trafficking.

In order to confirm the legitimate use of the OBU, it is necessary to determine a vehicle type, and for this purpose, a vehicle type discrimination apparatus capable of extracting information such as a vehicle width, a garage, and a vehicle head of a vehicle traveling through a road must be used.

Conventional vehicle type determination device for determining the type of the vehicle passing through the road and its operation will be described in detail by taking the Republic of Korea Patent Application 10-1997-0014458 as an example.

1 is a block diagram of a conventional vehicle type discrimination apparatus, and as shown in the related art, the vehicle type discrimination apparatus includes a vehicle detector 300, a vehicle separator 310, an axis number detector 320, an image photographing device 330, and a vehicle model discriminator 340. It is configured to include).

The vehicle detector 300 generates a vehicle detection signal by detecting a vehicle entering a predetermined lane such as a toll gate of a toll road, and preferably, a microwave sensor may be used.

The vehicle diverter 310 is installed at both sides of the lane, and detects a time point at which the front end of the vehicle passes and a time point at which the rear end of the vehicle passes.

The vehicle separator 310 includes a light emitting element and a light receiving element, and detects an optical signal generated by the light emitting element at the light receiving element and sets a time point at which the light cannot be detected due to interference of the vehicle. The optical sensor may not use the optical signal generated by the light emitting device, but may use an optical sensor that detects the time when the light may be detected by the passing of the vehicle.

The bearing sensor 320 has four switch contacts installed on the bottom of the lane perpendicular to the lane direction, and generates a predetermined contact signal by pressing pressure as the wheels of the vehicle pass through each switch contact point. The vehicle type discriminator 340 calculates the number of axes by analyzing the generation order and the number of times of the contact signal generated at each switch contact point of the bearing sensor 320, and determines the forward / reverse of the vehicle.

The image photographing apparatus 330 is installed above the bearing detecting unit 320, and photographs a background of the vehicle passing by the vehicle detection signal as a reference image, and between the detection start point and the detection end point. When the first axis of the vehicle is detected by the axis sensor, the upper end of the vehicle is photographed as a vehicle image.

The vehicle type discriminator 340 is an information processing system having a processor, a memory, an input / output device, and a predetermined software. The vehicle type discriminator 340 calculates a vehicle width and a vehicle head based on a difference value between the reference image and the vehicle image, The number of axes and the forward / reverse states of the vehicle are determined based on the generation order and the number of times of the contact signal generated at 320, and the type of the vehicle is determined. More specifically, the vehicle type discriminator 340 includes an image processor 350, a vehicle model determiner 360, and a contact signal processor 370.

The image processor 350 receives the reference image and the vehicle image from the imager 330, obtains a difference value between the two images, and calculates a vehicle width and a vehicle length. 354, an image storage unit 352, a difference value image generation unit 356, and a vehicle width / deputy image generation unit 358.

The image acquisition unit 354 receives the reference image and the vehicle image from the image photographing unit 330, and transfers the reference image to the image storage unit 352, and the vehicle image is the difference value image generation unit 356. ) To deliver.

The image storage unit 352 is a component block for temporarily storing a reference image transferred from the image acquisition unit 354. The image storage unit 352 compares the received reference image with a previously stored reference image, and if there is a significant difference, it is considered that an unnecessary vehicle or other object image is included in the reference image. Ignore

The difference value image generator 356 generates a difference value image by calculating a change amount of the vehicle image received from the image acquisition unit 354 and a reference image temporarily stored in the image storage unit 352.

The vehicle width / deputy calculator 358 calculates the width and the length of the vehicle entering the lane using the difference image generated by the difference image generator 356.

Next, the contact signal processing unit 370 determines the number of shafts and the forward / reverse state of the vehicle based on the generation order and the number of occurrences of each contact signal generated at the four switch contacts included in the bearing sensor 320. Transfer to the determination unit 360.

The vehicle type discrimination unit 360 compares the vehicle width and vehicle head calculated by the image processing unit 350 and the number of axes of the vehicle determined by the contact signal processing unit 370 with vehicle type data previously provided for each vehicle type. Select the matching vehicle.

2 is a view illustrating a conventional vehicle type determining apparatus in consideration of an installation environment. In FIG. 4, reference numeral 400 denotes the vehicle detector 300 of FIG. 1, reference numeral 402 denotes the vehicle separator 310 of FIG. 1, reference numeral 404 denotes the bearing sensor 320 of FIG. 1, and reference numeral 406 denotes FIG. The image photographing apparatus 330 of FIG. 1 is displayed.

Next, the operation of the conventional vehicle type discrimination apparatus will be described in detail in light of the passage of time.

FIG. 3 is a flowchart illustrating a method of determining a vehicle model in the conventional vehicle model determination apparatus shown in FIG. 1.

First, when a vehicle entering the lane is detected by the vehicle detectors 300 and 400, the background of the lane through which the vehicle passes by the image detectors 330 and 406 is photographed as a reference image, and the vehicle type discriminator 340. Are stored in a predetermined memory (steps 500 and 505). The process of storing the reference image in a predetermined memory will be described in more detail. The reference shape photographed by the image photographing apparatuses 330 and 406 and transmitted to the vehicle class discriminator 340 is called a first reference image. When the reference image already stored in the memory is referred to as the second reference image, the first reference image is ignored if there is a significant difference between the first reference image and the second reference image. The reference image is updated by the first reference image.

When the front end of the vehicle is detected by the vehicle separators 310 and 402 (step 510), and the first axis number of the vehicle is detected by the axis detectors 320 and 404 (step 515), the image photographing apparatus 330 In step 520, the vehicle image of the vehicle is captured.

Next, the difference between the photographed vehicle image and the reference image stored in the memory is calculated by the vehicle type discriminator 340 to calculate the vehicle width and the vehicle length (step 525).

Next, when the rear end of the vehicle is detected by the vehicle separators 310 and 402, the number of axes of the vehicle is determined based on the number of axes detected by the vehicle type detector 340 by the vehicle type discriminator 340. (Steps 530 to 545).

In addition, the vehicle type discriminator 340 compares the vehicle width, the vehicle head, and the number of vehicles with previously stored vehicle type data, and selects a matching vehicle type, thereby completing the vehicle type determination process for the vehicle entering the lane (step 550).

However, a method of acquiring a background image by a vehicle detector in advance when the vehicle enters a toll gate and acquiring an image of the vehicle when the vehicle enters a shooting area, and then obtaining a vehicle image using the background image and the image of the vehicle When the vehicle slows down with a delay, the background image may not be updated smoothly.

That is, although the entry of the vehicle is detected by the vehicle detector, not only the background is located in the current photographing area, but the previously entered vehicle may be located.

In addition, after acquiring the background image, there may be a difference between the background of the vehicle image and the previously acquired background image according to a change in conditions such as a cloud covering the sun. In this case, the vehicle image is used to It is impossible to determine.

As described above, the conventional vehicle type determination device cannot capture a desired background image when the vehicle is delayed or congested, and thus, it is impossible to discriminate the vehicle. there was.

In view of the above problems, the present invention can determine the type of vehicle irrespective of the delay and congestion of the vehicle, and has an object to provide a vehicle type determination apparatus and method irrespective of changes in weather.

1 is a block diagram of a conventional vehicle type discrimination apparatus.

FIG. 2 is a configuration diagram illustrating FIG. 1 in consideration of an installation environment; FIG.

3 is a flowchart of a conventional vehicle model discrimination method.

4 is a configuration diagram of a vehicle model discrimination apparatus of the present invention.

5 is an operational flowchart of FIG. 4;

Figure 6 is a flow chart of the vehicle model determination method of the present invention.

7A to 7H are schematic views of a process of extracting contours from an image of a vehicle.

8 is a schematic diagram showing a process of correcting and thinning the contour of the detected vehicle;

* Description of the symbols for the main parts of the drawings *

1: vehicle 2: vehicle detector

3, 4: first and second image capturing unit 5: image processing unit

6: Gantry

The above object is achieved by acquiring the width, height, and length information of the vehicle using only the image of the vehicle itself without using a background image. The present invention will be described in detail with reference to the accompanying drawings. Same as

4 is a block diagram of a vehicle type discrimination apparatus according to the present invention, and a vehicle detector 2 for detecting a position of a vehicle 1 traveling as shown therein; First and second imaging units 3 and 4 for imaging the upper and side portions of the vehicle 1, respectively, when the vehicle 1 passes through the vehicle detector 2; It is composed of an image processing unit (5) for receiving the image of the vehicle (1) taken by the first and second image pickup unit (3, 4) to process the vehicle, extract the height, width, and length information of the vehicle to determine the vehicle type .

Hereinafter, the configuration and operation of the vehicle type determination device of the present invention configured as described above will be described in more detail.

First, a vehicle detector 2 capable of detecting the passing of the vehicle 1 in a direction perpendicular to the road is provided.

The vehicle detector 2 has a sensor that operates by a change in pressure generated by the passing of the vehicle, that is, the weight of the vehicle, and the type thereof is not limited.

In addition, in order to capture the image of the side and the upper part of the vehicle 1 passing through the vehicle detector 2, a gantry 6 is provided, and the first and the second gantry 6 are disposed at appropriate positions of the gantry 6. The imaging units 3 and 4 are provided respectively.

The vehicle detector 2 plays a role of a switch for determining an imaging time point of the first and second imaging units 3 and 4. That is, the detection signal of the vehicle detector 2 that detects the vehicle 1 at the moment when the vehicle 1 passes the vehicle detector 2 is transmitted as an operation signal of the first and second imaging units 3 and 4. To perform the imaging operation.

Each of the first and second image capturing units 3 and 4 includes a CCD camera, an infrared filter, and an infrared light, and removes visible light in the imaging area due to the use of the infrared filter and the infrared light. Only with light.

The infrared filter is attached to the front surface of the lens of the CCD camera, and is made of glass or gelatin to apply only the wavelength of light that can be transmitted to the CCD camera to change the color tone or contrast of the subject.

When the upper and side surfaces of the vehicle are photographed using infrared rays, it is possible to solve an image blur problem caused by sunlight or the headlight of the vehicle.

As a result, it is possible to reduce the sensitivity of light, which is the biggest problem of image processing, and thus obtain an image irrespective of changes in weather.

The first and second image capturing units 3 and 4 are arranged to photograph the upper side and the side surface of the vehicle, respectively, thereby acquiring photographs photographed from different angles with respect to the same vehicle.

The width of the vehicle can be detected by using the imaging result of the first imaging unit 3, which photographs the upper part of the vehicle, and the imaging result of the second imaging unit 4, which photographs the side of the vehicle, is used. Since the length and height can be detected, the type of the vehicle can be determined by detecting all the width, the length, and the height of the vehicle.

The image of the vehicle 1 captured as described above is applied to the image processor 5.

The image processor 5 receives the image of the vehicle from the first and second image pickup units 3 and 4, and the image processor 5 processes the image of the vehicle to capture the width, length, and height of the vehicle. Use to determine the type of vehicle, and outputs and stores it.

FIG. 5 is an overall operation flowchart of the vehicle type discrimination apparatus according to the present invention. When the vehicle 1 is detected and the vehicle is detected, the first and second imaging units 3 and 4 capture images of the vehicle. Then, the captured image is processed to determine the vehicle type, and the result is output and stored.

In this process, the vehicle model can be directly determined by photographing the image of the vehicle passing through the toll gate without taking a separate background image as in the prior art, even when the vehicle is congested or delayed. It is possible to determine, and accurate vehicle discrimination is possible regardless of the change of weather.

Fig. 6 is a flowchart of an image processing method of the image processing section 5, and as shown therein, an image input step of reading images captured by the first and second image pickup sections 3 and 4; A preprocessing step of removing a background from the read image; A contour detection filtering step of detecting a contour of the vehicle from the image from which the background is removed; Thinning and contour detecting step of thinning the detected contour and determining the contour of the vehicle from the thinned contour; A contour candidate determining step of determining a candidate region of the vehicle contour using the detected contour; A contour candidate redetection step of re-detecting the contour of the contour candidate region of the vehicle; And determining a vehicle type by determining a length, a height, and a width of the vehicle from the outline of the vehicle.

Hereinafter, the image processing method of the image processing unit 5 as described above will be described in more detail.

First, in the image input step, images of the vehicles captured by the first and second image capturing units 3 and 4 are received.

FIG. 7A is a side image of the vehicle picked up by the second image pickup unit 4, and the processing of the upper image of the vehicle picked up by the first image pickup unit is the same as the side image processing process. Omit.

The side image of the vehicle captured by the second imaging unit 4 includes not only the vehicle but also the background. In order to detect only the outline of the vehicle from the image, the image of the vehicle is conventionally obtained by obtaining a difference between the two images. Although the contour is detected, in the present invention, the outline of the vehicle is detected by blurring the information of the image.

That is, FIG. 7B is a result of processing the side image of the vehicle illustrated in FIG. 7A by using a Gaussian filter, so that the outline of the vehicle can be detected by being blurred.

This is applied to the entire captured image using a 7 × 7 Gaussian filter in the preprocessing step of FIG. 6 to blur the entire image.

An example of the Gaussian filter is shown in Equation 1 below.

When a mask operation is performed using such a Gaussian filter, a point in the image is replaced by an average of values within a distance of three points based on the point, thereby blurring the image as a whole.

Then, outline outlines are extracted from the image shown in Fig. 7B.

In this case, we use a sobel filter to extract the edges of the image.

At this time, an example of a sobel filter is expressed as in Equation 2.

The filter expressed in Equation 2 has a large value when there are many changes in the upper and lower level values in the X axis based on one pixel, and when there are many changes in the left and right level values in the Y axis.

In other words, the value of X and Y increases at the portion that is the boundary, and the value decreases at the portion other than the boundary.

The size and direction of the edge detected using the filter can be expressed by the following equations (3) and (4).

The outline of the outline of the vehicle obtained as a result of the above detection is shown in Fig. 7C.

Next, contour detection filtering is performed, and contour detection at this time detects the contour by binarizing the preprocessed image as shown in FIG. 7C.

In the two-dimensional image, each pixel may be represented by f (x, y), and a threshold value T may be determined to compare the size of each pixel with a threshold value and convert the image into a value of 0 or 1. FIG.

Equation 5 below compares each pixel with a threshold to obtain g (x, y), which is a binarized result.

Through the above operation, one pixel has a value of 1 when the value is larger than T, and a value of 0 when the value is smaller than T, and the entire image becomes a binary image represented by two values of 0 or 1.

This is shown in Fig. 7d.

However, in the binary image, the contour of the vehicle may be distorted due to the noise component, or the contour may not be continued.

To solve this problem, the noise region is removed and the thinning and contour detection steps are performed to detect contour candidates connected by one line.

The thinning and edge detection steps first use dialing and erosion techniques to remove noise.

The dialation technique is the basic operation of mathematical morphology, which brings about the expansion effect of the object by the constituent element, which is shown in Equation 6.

Where A and B are sets and 0 is an empty set, the dialing of A by B is A It is represented by B. A is an input image, B is a mask operation element used for dialing, and B ^r is a reflection of B.

Equation (6) means that the dialing of A by B means all the points where the set x rotated around B and A overlap.

Through this process, the parts whose contours are not connected by the noise in the preprocessing process are connected.

When the dialation is performed, the unconnected contour part is connected, but the contour is thickened. To solve this problem, the object is reduced by the component using the erosion technique, and the technique is shown in Equation (7).

Equation (7) is the erosion of A by B means all points included in A converted by x, that is, B.

As such, the erosion can return the thickened contours to their original state.

Fig. 8A shows part A of a pre-processed image, Fig. 8B shows B, Fig. 8C shows an expanded area after dialing, and Fig. 8D shows a reduced area after erosion. .

B is an element that influences the result of each process, and determines the amount and shape of the area A to be expanded or reduced according to its size and shape.

As shown in FIG. 7E, the thinning and contour detection steps are performed as described above.

Since the contour of the vehicle extracted as described above is a contour detected from an image blurring by the first Gaussian filter, directly determining the vehicle model using the extracted Gaussian filter may reduce accuracy and reliability.

To solve this problem, the contour candidate detection step is extended by a variable distance D.

An enlarged outline is shown in FIG. 7F. In this case, the distance D used for extension is a value that is variably determined according to the accuracy of the boundary.

Then, as shown in Fig. 7G, the image of the vehicle originally photographed is compared with the boundary of the extracted and expanded vehicle.

As a result of the comparison, the vehicle model is determined using the extracted outline of the vehicle.

7H shows the contour of the vehicle finally extracted.

7A to 7H illustrate the image of the second image pickup unit 4, and the upper portion of the vehicle captured by the first image capture unit 3 is processed in the same manner as described above. The outline can be extracted.

Then, the width, length and height of the vehicle are detected using the extracted side profile of the vehicle and the top profile of the vehicle to determine the vehicle type of the imaged vehicle.

This determines the minimum rectangular area including the outline of the vehicle from the outline determined from the upper image of the vehicle captured by the first imager 3, and measures the length and width of the vehicle by measuring the length of the rectangle. do.

In addition, in the side image of the vehicle captured by the second imaging unit 4, the minimum rectangular area including the determined contour is determined, and the length and height of the vehicle can be known by measuring the length of the rectangle. .

At this time, the length of the vehicle can be detected by both the first imaging unit 3 and the second imaging unit 4, and the error can be known using the same.

As described above, the vehicle type discrimination apparatus of the present invention is configured to capture the side and the upper portion of the vehicle at the moment when the vehicle passes the toll gate, and to determine the vehicle type using only the captured image without comparing with other images. Alternatively, by not using the background information and the effect of determining the vehicle type of the vehicle running regardless of congestion, there is an effect of determining the correct type of vehicle regardless of the change of weather.

In addition, it is possible to extract the upper contour of the vehicle and the side contour of the vehicle by acquiring images of the upper and side surfaces of the vehicle, thereby making it possible to determine the vehicle type in consideration of the error, thereby improving the reliability of the vehicle type determination.

Claims (9)

A vehicle detector for detecting entry of the vehicle, and first and second imaging units for capturing images of the upper and side surfaces of the vehicle entering the vehicle, respectively; And an image processor for extracting the upper and side contours of the vehicle from the images captured by the first and second image capturing units, and measuring the height, length, width, and error of the vehicle from the contours. Vehicle model discrimination apparatus to do. The apparatus of claim 1, wherein the first and second imaging units comprise: an infrared illuminating unit for irradiating infrared rays; An infrared lens for selectively transmitting only the reflected light of the irradiated infrared rays; And a camera for capturing an image with infrared light incident through the infrared lens. The vehicle type discrimination apparatus according to claim 1, wherein the vehicle detector outputs a signal for driving the first and second image capturing units at the moment of detecting a vehicle that is installed in a direction perpendicular to a road. The image processing apparatus according to claim 1, wherein the image processing unit extracts an outline from the input image itself, corrects the outline, does not use the background image from the images captured by the first and second imaging units, and corrects the outline and the original. A vehicle type discrimination apparatus characterized by obtaining a contour having high reliability by comparing images. Imaging the top and side of the vehicle; Filtering the captured image to represent the blurred image; Extracting a contour candidate from the blurred image; Correcting and thinning the extracted contour candidates; Expanding the processed contour candidate and comparing the expanded contour candidate with the original image to determine a final contour if the contour candidate is appropriate; Calculating a width, a height, and a length of the vehicle using the outline, and determining a vehicle type by comparing with a reference value. The method of claim 5, wherein the blurring of the image comprises performing a blurring operation on the entire image. The method of claim 5, wherein the extracting the contour candidate comprises detecting a corner of each pixel of the blurred image, and extracting the contour candidate by binarizing an image including the detected corner information. Way. The method of claim 5, wherein the step of correcting and thinning the contour candidate is performed by enlarging and extending the contour of a portion that is not contoured by noise, and then reducing the contour to its original size. Determination method. The method of claim 5, wherein the determining of the vehicle type comprises calculating an error of vehicle type determination by using length information of the vehicle obtained from the upper outline of the vehicle and the side outline of the vehicle.