JPH0954892A - Vehicle width measuring method and device - Google Patents

Abstract

(57) [Summary] [Purpose] To provide a vehicle width measuring method and device which can accurately measure the vehicle width regardless of the vehicle body color and its density and even when the vehicle obliquely passes through the measuring section. To do. A measurement floor surface 22 having a striped pattern linearly extending in the width direction of the vehicle, a CCD camera 24 installed vertically downward on the top of the measurement floor surface, and an image processing device 26 for processing a captured image. (A) The CCD camera captures images of both sides of the vehicle against the measured floor surface, and (C) the difference between the vehicle density and the stripe density on the image is differentiated to perform a sudden change in density. A line segment is extracted. (D) Of the sudden change line segments, a line segment that extends for a predetermined length or more in the front-rear direction and is spaced most in the width direction is set as an extraction segment, and (E) is positioned on substantially the same line. The extracted segments are connected to obtain a pair of parallel lines, and the interval between the parallel lines is defined as the vehicle width.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle width measuring method and apparatus for automatically measuring the vehicle width of a traveling vehicle.

[0002]

2. Description of the Related Art Conventionally, it has been common to classify vehicle dimensions (vehicle length, vehicle height, vehicle width, etc.) by visual inspection in highways and parking facilities. However, this method has a problem that the accuracy is low and only rough classification is possible. In addition, the Ministry of Construction and the National Police Agency restrict the passage of special vehicles for road maintenance, etc., and have proposed an "automatic vehicle specification measurement device" for automatically identifying such special vehicles (for example, , JP-A-4-270497).

The device disclosed in Japanese Patent Laid-Open No. 4-270497 is shown in FIG.
As shown in FIG. 3, the ultrasonic heads 6a to 6h, the CCD cameras 7a to 7h, the light projectors 8 and 10, and the light receivers 9 and 11 are provided.
The vehicle height is measured by blocking the light between the light projector 10 and the light receiver 11, and the front end of the vehicle is detected by blocking the light between the light projector 8 and the light receiver 9. At that time, the CCD camera 7a ...
The rear end side of the vehicle is photographed by 7h, the edge detection processing is performed on the image data, the rear end and the left and right edges of the vehicle are detected, and the vehicle height at the CCD camera position is obtained by the ultrasonic heads 6a to 6h. , Correct the edge detection position error,
The vehicle width and vehicle length are calculated together with the installation position data.

[0004]

However, such a conventional automatic measuring device has the following problems. The detection accuracy of the ultrasonic sensor is low, so it is possible to roughly classify ordinary vehicles, large vehicles, special vehicles, etc. on highways etc., but measure the vehicle dimensions with a high accuracy of about several mm to determine the vehicle dimensions. The accuracy is too low to guide to the parking place (parking pallet etc.) provided for each of them, so it cannot be applied. Even if a CCD camera is used, when the vehicle body color is the same as the background color, extraction by image processing is almost impossible. When the vehicle passes the measuring section diagonally, automatic measurement is often impossible, and even if it is possible, an accurate vehicle width cannot be obtained.

Due to the above-mentioned various problems, in the conventional automatic vehicle width measurement, various kinds of vehicles (mainly passenger cars) constantly entering a large-scale (for example, 500 or more) mechanical parking device are continuously traveling. There was a problem that we could not classify accurately.

The present invention has been made to solve such a problem. That is, an object of the present invention is to provide a vehicle width measuring method and apparatus capable of instantaneously and accurately measuring the vehicle width of a vehicle when passing through the vehicle, thereby accurately classifying various vehicles having similar vehicle dimensions. To provide. Another object of the present invention is to reliably measure the maximum width of the vehicle regardless of the vehicle body color and its density, and to ensure the accurate vehicle width even when the vehicle diagonally passes through the measuring section. An object of the present invention is to provide a vehicle width measuring method and device capable of measuring.

[0007]

According to the present invention, there is provided a measurement floor having a striped pattern extending linearly in the width direction of the vehicle, and a CCD camera installed vertically downward on the measurement floor. With a CCD camera, images of both sides of the vehicle with the measurement floor surface as the background are captured, and the difference between the vehicle density on the image and the density of the striped pattern is differentiated to extract a sudden change line segment of the density, Of the sudden change line segments, a line segment that extends for a predetermined length or more in the front-rear direction and is spaced most in the width direction is used as an extraction segment, and extraction segments located on substantially the same line are connected to each other to form a pair of parallel lines. Is obtained and the interval between the parallel lines is set as the vehicle width, and a vehicle width measuring method is provided.
In addition, the measurement floor having a striped pattern extending linearly in the width direction of the vehicle, and a CCD installed vertically downward on the measurement floor and capturing images of both sides of the vehicle with the measurement floor as the background
The image processing device includes a camera and an image processing device that processes a captured image. The image processing device differentially processes the difference between the vehicle density on the image and the density of a striped pattern to extract a sudden change line segment of the density, Of the line segments, extend in the front-back direction for a specified length or more,
In addition, the line segment that is most distant in the width direction is set as the extraction segment, and the extraction segments that are almost on the same line are connected to each other.
There is provided a vehicle width measuring device characterized in that a pair of parallel lines is obtained, and an interval between the parallel lines is set as a vehicle width.

According to this method and apparatus, since the measurement floor surface having a striped pattern is imaged with the background as a background, by comparing at least one of the dark-colored portion and the light-colored portion of the striped pattern, regardless of the vehicle body color or its density. As a result, it is possible to reliably extract both side surfaces of the vehicle as a sudden change in density. Also, among the sudden change line segments, the line segment that extends for a predetermined length or more in the front-rear direction and that is most spaced in the width direction is used as the extraction segment, so a partially protruding portion such as a mirror is recognized as the vehicle width. It is possible to detect the accurate vehicle width without performing. Furthermore, since a pair of parallel lines is obtained by connecting the extraction segments located on substantially the same line and the interval between the parallel lines is taken as the vehicle width, even if the vehicle passes diagonally through the measuring section, the accurate vehicle width can be obtained. It is possible to measure reliably.

According to a preferred embodiment of the present invention, a vehicle detection sensor is installed in the passage area of the vehicle, and when the vehicle is detected by the sensor, images of both side surfaces of the vehicle are taken by the CCD camera. With this method, both side images of the central portion of the vehicle can be reliably captured from above. Further, a vehicle height measuring device is further provided to correct the vehicle width from the CCD camera and the vehicle height. By this correction, it is possible to eliminate the measurement error due to the distance between the CCD camera and the upper part of the vehicle.

It is preferable that the striped pattern is composed of dark color portions and light color portions of the same width, which are arranged alternately. The shaded portion may be black and white, or may be a different color.
With this configuration, it is possible to reliably extract both side surfaces of the vehicle as a sudden change line segment of the density regardless of the vehicle body color and the density thereof. Further, it is preferable that the CCD camera is composed of a pair of CCD cameras arranged at an interval in the width direction and capturing images simultaneously. With this configuration, the error due to the position of the CCD camera can be reduced, and the vehicle width can be accurately measured by the shading process regardless of the vehicle body color.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings. In the drawings, common parts are denoted by the same reference numerals. FIG. 1 is an overall configuration diagram of a vehicle width measuring device according to the present invention. In this figure, a vehicle width measuring device 20 according to the present invention has a measurement floor surface 22, C
A CD camera 24 and an image processing device 26 are provided. Further, in this figure, a vehicle detection sensor 27 is installed in a passage area of the vehicle 12, and when the vehicle 12 is detected by the sensor 27, images of both side surfaces of the vehicle 12 are included by the CCD camera 24 from above to the floor surface. To capture images.

The CCD camera 24 is installed vertically downward on the measurement floor surface 22 so as to capture images of both side surfaces of the vehicle 12 with the measurement floor surface 22 as a background from above. In this embodiment, the CCD cameras 24 are a pair of monochrome CCD cameras which are arranged at intervals in the width direction and are adapted to pick up images simultaneously. In this embodiment, two cameras are installed, one on the left and one on the right, but one may be installed if the vehicle width measurement accuracy may be rough. Further, when it is necessary to measure with higher accuracy, three or more units may be used together. Although a monochrome CCD camera is used here, a color CCD camera may be used. When using a plurality of cameras, all cameras simultaneously capture images in order to eliminate a shift due to the time of capturing images between the cameras.

FIG. 2 is a plan view of the measurement floor surface 22. As shown in this figure, the measurement floor surface 22 has a striped pattern extending linearly in the width direction of the vehicle 12. In addition, the striped pattern has dark color portions 22a and light color portions 2 of the same width which are alternately arranged.
2b. That is, since the body color has various shades from pure white to pure black, and has various colors,
Corresponding to these vehicles, the floor surface of the vehicle passage area has a certain width as shown in FIG. 2 so that there is always a clear difference in shade between the vehicle body color and the floor surface and both sides can be identified. Color part 2
2a (for example, black) and the light color portion 22b (for example, white) are alternately striped. Thus, by comparing at least one of the striped dark color portion 22a and the light color portion 22b, it is possible to reliably extract both side surfaces of the vehicle as a sudden change line segment of the density regardless of the vehicle body color and the density thereof.

The widths of the dark color portion 22a and the light color portion 22b are preferably made sufficiently longer than that of the mirror or the like in order to distinguish the width from the protrusion of the mirror or the like. The floor light and shade colors are not limited to white and black as long as the light and shade colors are clearly separated to some extent.

FIG. 3 is an image processing flowchart in the image processing device 26, and FIG. 4 is a schematic diagram of this image processing. The image processing device 26 processes the captured image captured by the CCD camera 24 and outputs the vehicle width. That is, the image processing device 26 differentiates the difference between the vehicle density on the image and the density of the striped pattern to extract a sudden change line segment of the density (FIG. 4C).
A line segment that extends substantially a predetermined length or more in the front-rear direction and is most spaced in the width direction is used as an extraction segment (FIG. 4D), and extraction segments located on substantially the same line are connected to obtain a pair of parallel lines. The distance between these parallel lines is the vehicle width (Fig. 4E).
It has become.

Hereinafter, FIG. 3 will be described in more detail. Image input (imaging) By the trigger signal from the vehicle detection sensor 27 described above,
Images are simultaneously input from the left and right CCD cameras 24. Take out the parts including the left and right sides of the car body from each image, 1
Combine into one image. Note that here, two images are
Although there are two images, the following processing may be performed on each of the right and left sides instead of the single image. Further, noise processing such as smoothing processing is performed on the input image as necessary.

Creation of Edge Strength Image / Direction Image An "edge strength image" showing the strength of the differential value (strength of gray level change) by differentiating the gray level of the image and the direction of differentiation (of the gray level change) "Edge direction image" indicating the direction) is created. Further, with respect to the “edge strength image”, a threshold value is set in advance for the strength, and the threshold value is binarized. Furthermore, similarly for the "edge direction image",
A threshold value is set for the direction (angle / direction corresponding to the passing direction of the vehicle), and the threshold value is binarized.

Edge Image Creation Logical product (A of edge strength image) and edge direction image (A
ND). That is, an edge having a certain degree of edge strength and extending in the vehicle traveling direction is extracted. Noise processing Removes edges that have an area (line segment of length) that is less than a certain value.

Polygonal Approximation of Edges Polygonal approximation is applied to the edges remaining by the noise processing. This processing is performed because the remaining edges are not always straight lines (circular arcs and broken lines are displayed). When executing the polygonal line approximation, the straight line component below the threshold is deleted. The edges remaining in the approximation are considered to be straight line segments.

Selection of Segments at Left and Right Ends Segments at left and right ends in the image, that is, segments at both left and right ends of the vehicle are selected from the segments approximated to the broken line. Each segment is a linear expression ax + by + c = 0. . . Since it is expressed by Equation 1, from the coefficients (a, b, c) of the equation,
Select the left and right segments.

Integration of Segments The left and right ends of each segment are integrated with the remaining segments. That is, the coefficient comparison of Expression 1 and the value (y1) of the point (x1, y1) on the segment to be compared are substituted into the expression at the left (right) end, and x obtained from that is substituted.
And the value of x1 are compared, and if they are equal to or less than the threshold value, they are regarded as the same linear expression (integrated). This test is performed on all remaining segments.

Linear fitting For the set of each segment at the left and right ends that have been integrated,
Apply the linear formula. The linear fitting is performed by, for example, the least squares approximation method. Further, the reference points when using the least-squares approximation method are both end points of each segment. At this time, the height is corrected using the vehicle height information. That is,
The higher the vehicle height is with respect to the floor surface, the shorter the vehicle width is measured. Therefore, the vehicle height value measured separately is used to correct the height direction.

A linear distance between the left and right ends obtained by the calculation of the vehicle width, that is, the vehicle width is calculated.

Next, FIG. 4 will be described in comparison with FIG. In FIG. 4A, the CCD camera captures images of both sides of the vehicle with the measurement floor surface as the background, and in FIG. 4B, the left and right images are combined. These processes correspond to those in FIG. Next, in FIG. 4C, the difference between the vehicle density on the image and the density of the striped pattern is differentiated to extract a sudden change line segment of the density. This process corresponds to that in FIG.

Next, in FIG. 4 (D), among the abruptly changing line segments, the line segment extending in the front-rear direction by a predetermined length or more and having the most space in the width direction is set as the extraction segment. This process is equivalent to in FIG. Finally, FIG.
In (E), a pair of parallel lines is obtained by connecting the extraction segments located substantially on the same line, and the interval between the parallel lines is taken as the vehicle width. These processes correspond to in FIG. Further, it is preferable that a vehicle height measuring device is further provided to correct the vehicle width calculated from the CCD camera 24 and the vehicle height in FIG. 4 (E).

According to the method and apparatus of the present invention described above,
Since the measurement floor surface 22 having a striped pattern is imaged against the background, comparison with at least one of the dark color portion 22a and the light color portion 22b of the striped pattern indicates that the color is independent of the vehicle body color and its density.
Both side surfaces of the vehicle 12 can be reliably extracted as a sudden change in density. Also, among the sudden change line segments, the line segment that extends for a predetermined length or more in the front-rear direction and that is most spaced in the width direction is used as the extraction segment, so a partially protruding portion such as a mirror is recognized as the vehicle width. It is possible to detect the accurate vehicle width without performing. Furthermore, since a pair of parallel lines is obtained by connecting the extraction segments located on substantially the same line and the interval between the parallel lines is taken as the vehicle width, even if the vehicle passes diagonally through the measuring section, the accurate vehicle width can be obtained. It is possible to measure reliably.

[0027]

Embodiments FIGS. 5 to 7 are processed images displayed on a display showing the embodiment of the present invention described above. In each figure, (A) is a composite image, (B) is an image after image processing,
4A corresponds to FIG. 4B, and FIG. 4B corresponds to FIG. 4D.
Is equivalent to FIG. 5 shows the case where the vehicle body color is black or a dark color. From FIG. 5 (B), the light-colored portion 22b (white in this example) on the measurement side surface 22 has a sharp density change line segment extracted. Recognize. On the contrary, FIG. 6 shows the case where the vehicle body color is white or a light color, and the dark color portion 22a on the measurement side surface 22 is seen from FIG. 6B.
(In this example, black), the sudden change line segment of the density is extracted. Further, FIG. 7 shows a case where the vehicle passes diagonally,
Even in this case, it can be seen from FIG. 7C that the accurate vehicle width can be reliably measured.

Therefore, from these embodiments, the maximum width of the vehicle can be reliably measured regardless of the vehicle body color and its density, and the accurate vehicle width can be ensured even when the vehicle obliquely passes through the measuring section. You can see that it can be measured.

The present invention is not limited to the above-described embodiments, and it goes without saying that various modifications can be made without departing from the gist of the present invention.

[0030]

As described above, the vehicle width measuring method and apparatus according to the present invention can measure the vehicle width of a vehicle instantaneously and accurately when the vehicle passes, and thus various types of vehicles having similar vehicle dimensions can be measured. Can be classified accurately. Further, the maximum width of the vehicle can be reliably measured regardless of the vehicle body color and its density, and the vehicle width can be accurately measured even when the vehicle obliquely passes through the measuring section.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of a vehicle width measuring device according to the present invention.

FIG. 2 is a plan view of a measurement floor surface 22.

FIG. 3 is an image processing flowchart in the image processing device 26.

FIG. 4 is a schematic diagram of image processing.

FIG. 5 is a halftone image displayed on the display when the vehicle body color is dark.

FIG. 6 is a halftone image displayed on the display when the vehicle body color is light.

FIG. 7 is a halftone image displayed on the display when the vehicle passes diagonally.

FIG. 8 is a schematic configuration diagram of a conventional vehicle width measuring device.

[Explanation of symbols]

 6a to 6h Ultrasonic head 7a to 7h CCD camera 8,10 Light projector 9,11 Light receiver 12 Vehicle 20 Vehicle width measuring device 22 Measurement floor surface 22a Dark color portion 22b Light color portion 24 CCD camera 26 Image processing device 27 Vehicle detection sensor

Claims (6)

[Claims] 1. A measurement floor surface having a striped pattern linearly extending in the width direction of a vehicle, and a CCD camera installed vertically downward on an upper portion of the measurement floor surface. An image of both sides of the vehicle is taken against the background, and the difference between the vehicle density on the image and the density of the striped pattern is differentiated to extract a sudden change line segment of the density. A segment that extends for a predetermined length or more and that is most spaced in the width direction is used as an extraction segment, and extraction segments that are located substantially on the same line are connected to obtain a pair of parallel lines. A vehicle width measuring method characterized by being width. 2. A vehicle detection sensor is installed in a passage area of the vehicle, and when the vehicle is detected by the sensor, images of both side surfaces of the vehicle are picked up by a CCD camera. Vehicle width measurement method. 3. The vehicle width measuring method according to claim 1, further comprising a vehicle height measuring device, wherein the vehicle width is corrected based on the CCD camera and the vehicle height. 4. An image of a measurement floor having a striped pattern extending linearly in the width direction of the vehicle, and vertically downward on the measurement floor, and capturing images of both sides of the vehicle against the background of the measurement floor. A CCD camera and an image processing device for processing the picked-up image, and the image processing device differentiates the difference between the vehicle density on the image and the density of the striped pattern to extract a sudden change line segment of the density, Of the sudden change line segments, a line segment that extends substantially a predetermined length or more in the front-rear direction and is most spaced in the width direction is an extraction segment,
A vehicle width measuring device, characterized in that a pair of parallel lines is obtained by connecting extraction segments located on substantially the same line, and a space between the parallel lines is defined as a vehicle width. 5. The striped pattern is composed of dark color portions and light color portions having the same width and arranged alternately.
Vehicle width measuring device described in. 6. The vehicle width measuring device according to claim 4, wherein the CCD cameras are composed of a pair of CCD cameras arranged at intervals in the width direction and capturing images at the same time.