JP2002024809A - Vehicle detection apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a vehicle detection apparatus which functions accurately to enhance precision, capable of functioning accurately without being influenced by changes in sunshine conditions or the like, for detecting vehicles entering and leaving a parking lot, a tollgate for a toll road, a service area, a parking area or the like. SOLUTION: An image photographed at present time and a photographed image prior to the present in an earlier time are obtained by an imaging device capable of photographing a vehicle passing on a lane, continuously at intervals of a fixed time, and difference calculation of respective image data is executed by a difference operation circuit. Passage of the vehicle is decided based on the characteristic obtained from the result, to thereby reduce the difference between backgrounds of the photographed images, and to enable proper vehicle detection, without being influenced by changes in sunshine conditions or the like.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle detecting device for detecting vehicles entering and exiting a parking lot or a tollgate, a service area, or a parking area of a toll road.

[0002]

2. Description of the Related Art In a parking lot, a tollgate on a toll road, a service area, a parking area, or the like, it is detected that a vehicle has entered the gate, and an entrance / exit gate is opened. It is necessary to output information to a higher-level device, such as issuing a ticket or instructing the operation of a fee settlement machine.

[0003] As a vehicle detecting device for this purpose, there are a conventional system using an ultrasonic sensor, a system using an electromagnetic wave and a photoelectric sensor, and the like. Here, an image processing system will be described with reference to FIGS. Will be explained.

FIG. 7 is an explanatory view showing the overall structure of a conventional vehicle detecting device. FIG. 8 is an explanatory view for explaining the process of vehicle detection. FIG. 7 (a) is an input image, and FIG. 8 (b) is a reference image. , And (c) show the detection result.

[0005] Reference numeral 1 denotes an imaging device which is arranged in the direction of approach of a gate or a lane 12 connected to the gate, and captures an image of a vehicle 10 entering through the lane 12 and a background 11 connected behind the vehicle 10. .

Reference numeral 2 denotes an image memory which digitizes an image picked up by the image pickup device 1 and stores it as image data. Reference numeral 3 denotes a reference memory in which an image in which only the background 11 when the vehicle 10 does not exist is captured in advance is registered.

When an image is sent from the image pickup apparatus 1 to the image memory 2, the reference memory 3 also functions, and the contents of the image memory 2 and the reference memory 3 are input to the absolute difference operation circuit 21 and The absolute value of the difference is calculated for each pixel of the input image.

The output of the absolute difference calculation circuit 21 is input to the binarization circuit 6, and a region of a pixel whose absolute value of the difference is larger than a predetermined threshold value is detected. Is input to the feature extraction circuit 8 as the output of

The feature extraction circuit 8 measures and outputs the area output from the binarization circuit 6, that is, the area of the vehicle area, and a feature amount on an image such as a circumscribed rectangle. Is displayed on the display 9 or is output to a higher-level device for opening an entrance / exit gate, issuing a toll ticket, instructing the start of operation of a toll settlement machine, and the like by a communication means (not shown), and controlling the toll machine. Is performed.

[0010] This flow will be further described with reference to FIG.
When the image capturing apparatus 1 captures an image of the vehicle 10 that has entered as shown in FIG. 7, an image including the vehicle 10 and the background 11 is captured as an “input image” as shown in FIG. Is stored in the image memory 2.

On the other hand, since an image in which only the background 11 as shown in FIG. 8B is captured in the reference memory 3 as a "reference image" in advance, the "reference image" and the "input Is sent to the absolute difference calculation circuit 21 as image data, and the absolute difference calculation circuit 21 calculates the absolute value of the difference between the contents of the image memory 2 and the contents of the reference memory 3 for each pixel.

In the background 11 portion of the image, since the image data of both the input image and the reference image are the same image, the difference is calculated to be 0. The density of the pixel is changing, and some value is output when the difference is obtained.

Here, the sign of the difference value is reversed between the case where the paint color of the vehicle 10 captured in the input image is white and the case where the paint color is black. Only shows output above a certain value.

The output of the absolute difference calculation circuit 21 is compared with a preset threshold value in the binarization circuit 6,
Pixels above the threshold output "1", and pixels below the threshold output "0".

As a result, an image in which only the area of the vehicle 10 is "1" is obtained as a detection result as shown in FIG.
Next, the feature extraction circuit 8 obtains a feature amount of the vehicle 10 region as shown in FIG.

The feature amount is the vehicle 10 region in the height direction of the upper and lower ends j _min, j _max and shown in FIG. 8 (c), the horizontal direction of the left and right ends i _min, and the circumscribed quadrangle of the i _max, etc., area, i.e., “1”
Is used. The output of the feature extraction circuit 8 is displayed to the operator on the display 9 so that necessary actions such as opening and closing the gate are performed.

[0017]

However, in the above-described conventional apparatus, by using the absolute difference calculation circuit 21 for detecting the vehicle 10, an error due to a difference between the case where the paint color of the vehicle 10 is white and the case where the paint color of the vehicle 10 is black can be eliminated. Although it is considered that no occurrence occurs, when calculating the background difference, if the density of the entire image changes due to a change in the sunshine condition, etc., all the regions are determined to be changed regions, that is, the vehicle 10. Sometimes.

Further, in the vehicle detection apparatus using the background difference calculation circuit according to the above-mentioned prior art, when the density of the entire image changes due to a change in sunshine or the like, it is determined that the road surface reflection of the headlight at night indicates that the vehicle is present. Therefore, there is a problem that reliability is lowered.

The present invention has been made to solve the above-mentioned problems in the conventional apparatus, and to provide a suitable vehicle detection apparatus which functions more accurately and is more accurate without being affected by changes in the sunshine conditions. Is the subject.

[0020]

Means for Solving the Problems The present invention has been made to solve the above-mentioned problems. As a first means, the present invention is directed to a vehicle which is arranged toward a lane in which a vehicle passes, and in which a vehicle passing through the lane is fixed. An imaging device capable of continuously capturing images at each time interval; a first image memory for storing image data by digitizing a captured image of the imaging device at the current time; and capturing an image of the imaging device one time before the current time A second image memory for digitizing the image and storing the image data;
A difference calculation circuit for calculating a difference between the image data of the first and second image memories, and binarizing an output of the difference calculation circuit to extract a predetermined feature, and determining whether the vehicle has passed according to the feature; A vehicle detection device provided with a circuit is provided.

That is, according to the first means, of the images continuously taken by the imaging device, the image at the current time is stored in the first image memory, and the image one time before the current time is stored in the first image memory. Two image memories are stored, and the difference between the image data at the current time and the image data one time before the current time is calculated by a difference calculation circuit. Based on the characteristic obtained from the difference calculation, the passage of the vehicle is determined. So I do
The image data for which the difference is calculated is compared in substantially the same background without being shifted for only one time in captured images that are continuously captured, and is affected by changes in sunlight conditions and the like. Instead, it is possible to reliably detect an image change due to the entry of the vehicle and perform appropriate vehicle detection.

According to a second aspect of the present invention, in the first aspect, in the first aspect, the imaging device is configured such that an outer shape of a vehicle passing through a lane or headlights (in this specification, headlights and tail lamps are collectively referred to as headlights and tail lamps). And a headlight etc.), and processes the photographed image as the photographed image.

In other words, according to the second means, the photographed images continuously photographed by the image pickup device are formed in the shape of a vehicle passing through a lane, the headlights of the vehicle, or the shape of the vehicle and the head. Since it is both light and so on,
When the outer shape of the vehicle is taken as a photographed image, it is adapted to detection of a vehicle in daytime, and when it is used as a headlight of a vehicle, the outer shape of the vehicle is adapted to detection of a vehicle at night or the like where it is difficult to determine the outer shape of the vehicle, and When both the outer shape of the vehicle and the headlights are targeted, the vehicle can be detected at any time during 24 hours a day, and appropriate vehicle detection can be performed in accordance with each scene.

[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is an explanatory diagram showing an overall configuration of a vehicle detection device according to the present embodiment. FIG. 2 is an explanatory diagram for explaining a vehicle detection process by the vehicle detection device of FIG. The image, (b) shows the input image one time before the current time, and (c) shows the detection result.

In order to avoid redundant description, the same parts as those of the above-described conventional one are denoted by the same reference numerals in the drawings, and redundant description will be omitted as much as possible in this embodiment. We will focus on specific points.

That is, the present embodiment is employed mainly in the daytime, and functions more effectively in a situation in which the surrounding environment is bright and the entire appearance and overall shape of the approaching vehicle can be sufficiently grasped. It is a form of doing.

Reference numeral 1 denotes an image pickup device which is arranged toward a lane 12 through which the vehicle 10 passes, and
Can be continuously photographed at fixed time intervals, such as 1/30 second intervals or 1/15 second intervals depending on the situation.

Reference numeral 2a denotes a first image memory, 2b denotes a second image memory, the first image memory 2a captures an image at the current time t captured by the imaging device 1, and the second image memory 2b stores one image from the current time. The image taken before (t-1) is transmitted, digitized, and stored.

Numeral 22 denotes a difference operation circuit which connects the first and second image memories in parallel, and stores the image data at the current time t in the first image memory 2a and the second image memory 2
The image data one time before (t-1) before the current time in b is transmitted, and the difference between the two image data is calculated.

Reference numeral 23 denotes a binarizing circuit, which is a differential operation circuit 2
2 and binarizes the time difference image output from the difference calculation circuit 22.

Numeral 8 denotes a feature extracting circuit, 13 denotes a vehicle entry judging circuit, and the feature extracting circuit 8 comprises a binarizing circuit 23 on the upstream side.
And the downstream side is connected to the vehicle entry determination circuit 13 and
A feature such as area is calculated from the binary image output from the digitizing circuit 23, and the calculated feature is transmitted to the vehicle approach determination circuit 13, and the vehicle approach determination circuit 13 determines the vehicle approach. I have.

In the present embodiment configured as described above, the image picked up by the image pickup
(I, j, t), the time (t−
Assuming that the image captured in 1) is I (i, j, t−1), the difference calculation circuit 22 uses the difference image S (i, j, t).
= I (i, j, t) -I (i, j, t-1) (where I is the Image and S is the Subtract initial for convenience).

The difference image S generated as described above
(I, j, t) is converted to a threshold value T by the binarization circuit 23.
Binarization is performed using H1 and TH2 (TH1> TH2) to generate a binary difference image BD (i, j, t) (where, TH is an abbreviation for Threshold, and BD is a background day initial for convenience). Use). Here, binarization is the next operation.

If (S (i, j, t)> TH1 or S
(I, j, t) <TH2) BD (i, j, t) = 1 else BD (i, j, t) = 0.

Here, the binarized threshold values TH1 and TH2
Is given as a fixed value by the user, but there is also a method of dynamically obtaining the standard deviation based on the standard deviation on the road surface portion which is the background 11 (there is no vehicle).

Also, the difference in luminance of the difference image may be different between the case where the vehicle 10 is painted white and the case where the painted color of the vehicle 10 is black. Therefore, the white vehicle and the black vehicle can be obtained by using the two threshold values TH1 and TH2. Both vehicles can be extracted.

Next, BD (i,
j, t), and counts the number of pixels of
Accordingly, when the counted number of pixels (area) is equal to or larger than the threshold value, it is determined that the vehicle has entered.

Here, although the area is a feature that is effective for vehicle entry, when the area is counted, the image is meshed, the area count is performed for each mesh, and the vehicle entry is performed by relative comparison of the area for each mesh. You can also make decisions,
This makes it possible to distinguish between a large vehicle and a small vehicle.

As described above, according to the present embodiment, since the difference is calculated by performing the difference calculation on the image taken by the imaging apparatus 1 before and after one time, the feature is extracted. It works well and is especially useful for daytime vehicle detection.

Next, a second embodiment of the present invention will be described with reference to FIGS. 3 is an explanatory diagram showing the overall configuration of the vehicle detection device according to the present embodiment, FIG. 4 is an explanatory diagram of the operation, (a) illustrates the expansion operation by the maximum value filter, and (b) illustrates the label characteristics. The description of headlight extraction is shown.

In order to avoid redundant description, the same parts as those of the above-described conventional one and the first embodiment are denoted by the same reference numerals in the drawings, and redundant description is omitted as much as possible. The following description focuses on features specific to the present embodiment.

In the first embodiment, the vehicle is detected by identifying the outer shape of the vehicle as a moving body in the daytime, but the vehicle body cannot be seen in the nighttime image.
In the present embodiment, the vehicle is detected by extracting the headlights of the vehicle heading for the image pickup device or the tail lamp of the vehicle moving away from the image pickup device, and is used mainly at night to function effectively. .

In the present embodiment, the image pickup device 1 arranged toward the lane 12 through which the vehicle 10 passes has a structure capable of continuously taking images at a fixed time interval, and takes an image at the current time t in the first image memory. 2a, and the image taken one time before (t-1) before the current time is transmitted to the second image memory 2b and digitized, and the image data in the first and second image memories 2a and 2b is subtracted. The difference calculation is performed by the calculation circuit 22, and the time difference image output from the difference calculation circuit 22 is binarized by the binarization circuit 6. Except for being directed to the lights or the tail lamp of the vehicle, it is exactly the same as the first embodiment.

However, in the present embodiment, the binary image of the binarizing circuit 6 is transmitted to the maximum value filter circuit 24, and the expansion operation is performed on the binary image. The configuration is such that labeling processing and features such as area are calculated from the binary image output from the maximum value filter circuit 24, and finally, the vehicle approach determination circuit 14 determines the vehicle approach.

In the present embodiment configured as described above, the image picked up by the image pickup device 1 at time t is
(I, j, t), the time (t−
Assuming that the image captured in 1) is I (i, j, t−1), the difference calculation circuit 22 uses the difference image S (i, j, t).
= I (i, j, t) -I (i, j, t-1).

Images I (i, j, t) and I (i, j, t-
1) is on the image memory 2. First, a light candidate is extracted by a binarization process using the fact that the light portion has a higher luminance value than the background 11.

Next, the generated difference image S (i, j,
t) is binarized by the threshold value TH3 in the binarization circuit 6 to generate a binary difference image BN (i, j, t) (where BN is a background night initial for convenience). Use). Here, binarization is the next operation.

If (S (i, j, t)> TH3) BN
((I, j, t) = 1 else BN ((i, j, t) = 0.

Here, the binarization threshold TH3 is given as a fixed value by the user, but there is also a method of dynamically obtaining the threshold TH3 based on the average luminance and the standard deviation in the background.

The BN ((i, j, t)) generated by the binarization circuit 6 as described above is subjected to the maximum value filter circuit 2
4 is performed to prevent the label of the headlight candidate from being divided.

That is, since the generated image includes a road surface line that appears due to light irradiation and road surface reflection of the light in addition to the headlight candidates, in order to correctly detect the vehicle, the light and the light are required. Need to be distinguished from others.

In this identification processing, a label feature amount is used.
The binary image is input to the feature extraction circuit 8 and subjected to labeling processing to calculate label feature amounts such as height, width and area. While the label of the light has a shape close to a circle, the reflection of the light and the road surface line have a vertically long elliptical shape.

Therefore, a label whose label has a circumscribed rectangle with a vertical side length of a certain value or more or a label with a large aspect ratio can be eliminated as a disturbance, and noise has a small area, so that the label area is small. If a certain threshold is not reached, it is removed as noise. If there is a headlight candidate after noise removal, the vehicle entry determination circuit 14 determines that there is a vehicle entry.

Although the object to be photographed by the image pickup apparatus 1 has been described as the headlight of the entering vehicle,
Of course, this imaging may be a tail lamp of a vehicle that is leaving, and in that case, the headlight in the above description may be replaced with a tail lamp.

As described above, according to the present embodiment, the vehicle headlight (or the tail lamp of the vehicle moving away from the image pickup device) that is directed toward the image pickup device 1 is extracted to detect the vehicle, and is mainly used at night. Being informative.

Next, a third embodiment of the present invention will be described with reference to FIGS. FIG. 5 is an explanatory diagram showing the overall configuration of the vehicle detection device according to the present embodiment, and FIG. 6 is an explanatory diagram of the vehicle entry determination integrated circuit in FIG.

In order to avoid redundant description, the same parts as those of the above-described conventional one and the first and second embodiments are denoted by the same reference numerals in the drawings, and will be described repeatedly. Will be omitted as much as possible, and points specific to the present embodiment will be mainly described.

That is, at dusk, the contrast between the vehicle and the road surface is lower than in the daytime, so that the same detection method as in the daytime according to the first embodiment may cause the vehicle to miss detection. According to the second embodiment, the same detection method as at night may cause a detection failure of a vehicle whose light is not turned on. Therefore, in the present embodiment, the daytime and nighttime detection methods are parallelized. Vehicle detection is performed by using this function to make the vehicle effective at twilight.

Therefore, in the present embodiment, as is apparent from the description of FIG. 5, the first image memory 2a to which the imaging at the current time t is transmitted from the imaging device 1 and the time before the current time (t For the second image memory 2b to which the imaging of (-1) is transmitted, a series (hereinafter referred to as a first series for convenience) of the difference operation circuit 22 and below described as the first embodiment, A series (hereinafter referred to as a second series for convenience) connected in parallel with the difference calculation circuit 22 and the following series described in the form, and a vehicle entry determination integration circuit 15 for performing a comprehensive determination of the first and second series at the end thereof Are arranged.

The specific structure, operation, and the like of the first and second systems are the same as those described in the first and second embodiments. Here, a description will be given of the vehicle entry judgment integration circuit 15 combining the first and second streams.

That is, as shown generally in FIG. 6, in the vehicle entry judgment integration circuit 15, the result of the first series corresponding to the daytime vehicle entry judgment or the second series of the result corresponding to the nighttime vehicle entry judgment is obtained. If at least one of the results is determined to be "entered", the determination of "entered" is made as an integrated determination result.

Thus, according to the present embodiment, the approach of the vehicle can be surely grasped even during the twilight between daytime and nighttime, and vehicle detection can be appropriately performed over 24 hours a day.

Although the present invention has been described with reference to the illustrated embodiment, the present invention is not limited to such an embodiment.
It goes without saying that various changes may be made to the specific structure within the scope of the present invention.

[0064]

As described above, according to the first aspect of the present invention, the vehicle is arranged toward the lane where the vehicle passes, and the vehicle passing through the lane can be continuously photographed at regular time intervals. An imaging device, a first image memory for digitizing a captured image of the imaging device at the current time and storing image data, and digitizing a captured image of the imaging device one time before the current time and storing the image data A second image memory, a difference calculation circuit for calculating a difference between the image data of the first and second image memories, and binarization of an output of the difference calculation circuit to extract a predetermined feature; Since the vehicle detection device is configured by providing a circuit for determining the passage of the vehicle, the captured image at the current time stored in the first image memory among the captured images continuously captured by the imaging device, Record in image memory The difference calculation circuit calculates the difference of the photographed image one time before the current time, and determines whether the vehicle has passed based on the feature obtained from the difference calculation. Of the consecutively photographed images, only one time interval is compared, and the comparison is performed with substantially the same background, and the image change due to the entry of the vehicle is not affected by the change of the sunshine condition and the like. , And appropriate vehicle detection is performed. In particular, a suitable vehicle detection device used in the daytime could be obtained.

According to a second aspect of the present invention, in the first aspect of the present invention, the imaging device captures at least one of an outer shape of a vehicle passing through a lane, a headlight, and the like. Since the vehicle detection device is configured to process the image as the photographed image, the photographed image continuously photographed by the image pickup device may be an external shape of a vehicle passing through a lane or a headlight of the vehicle. By using both the outer shape of the vehicle and the headlights, etc., when the outer shape of the vehicle is used as a photographed image, it is adapted to detection of the vehicle in the daytime, and when the outer shape of the vehicle is used as the headlights, the outer shape of the vehicle is used. Adapts to the detection of vehicles at night or the like when it is difficult to determine, and also covers both daytime and nighttime twilight when targeting both the outer shape and headlights of the vehicle, Day even for the case of what time in the 24-hour in which it was possible to obtain a suitable vehicle detection device can be adapted.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram illustrating an overall configuration of a vehicle detection device according to a first embodiment of the present invention.

FIGS. 2A and 2B are explanatory diagrams illustrating a process of vehicle detection by the vehicle detection device of FIG. 1, wherein FIG. 2A is an input image at a current time, FIG. 2B is an input image one time before the current time, and FIG. Indicates a detection result.

FIG. 3 is an explanatory diagram illustrating an overall configuration of a vehicle detection device according to a second embodiment of the present invention.

4A and 4B are explanatory diagrams of the operation of the apparatus of FIG. 3, wherein FIG. 4A illustrates the expansion operation by a maximum value filter, and FIG. 4B illustrates the headlight extraction by a label feature.

FIG. 5 is an explanatory diagram showing an overall configuration of a vehicle detection device according to a third embodiment of the present invention.

FIG. 6 is an explanatory diagram of a vehicle entry judgment integrated circuit in FIG. 5;

FIG. 7 is an explanatory diagram showing an overall configuration of a conventional vehicle detection device.

FIG. 8 is an explanatory diagram illustrating a process of vehicle detection in FIG. 7;
(A) shows the input image, (b) shows the reference image, and (c) shows the detection result.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 imaging device 2 image memory 2a image memory 2b image memory 3 reference memory 6 binarization circuit 8 feature extraction circuit 9 display 10 vehicle 11 background 12 lane 13 vehicle approach determination circuit 14 vehicle approach determination circuit 15 vehicle approach determination integrated circuit 21 Absolute difference calculation circuit 22 Difference calculation circuit 23 Binarization circuit 24 Maximum value filter circuit

Claims (2)

[Claims] Claims: 1. The vehicle is arranged toward a lane through which a vehicle passes,
An imaging device capable of continuously capturing a vehicle passing through the lane at regular time intervals; a first image memory for digitizing a captured image of the imaging device at the current time and storing image data; A second method of digitizing a captured image of the imaging device before time and storing image data
An image memory, a difference calculation circuit for calculating a difference between the image data of the first and second image memories, and a predetermined feature extracted by binarizing an output of the difference calculation circuit. A vehicle detection device comprising a circuit for judging passage. 2. The vehicle according to claim 1, wherein the imaging device captures an image of at least one of an outer shape of a vehicle passing through a lane, a headlight, and the like, and processes the captured image as the captured image. Detection device.