JPH0996528A - Device and method for detecting distance between vehicles - Google Patents

Abstract

PROBLEM TO BE SOLVED: To measure the distance to a preceding vehicle by photographing two high luminant shapes being provided in the rear of the preceding vehicle. SOLUTION: Two square-shaped infrared ray generators 12 and 14 are provided in the rear of a preceding vehicle 100, and a CCD camera 16 of a following vehicle 200 is used for shooting. The obtained image is supplied to an image processer ECU 18. The image processer ECU 18 extracts one square high luminant region from the image. The position of another high luminant region is estimated from the size of the extracted high luminant region. When a high luminant image exists in the estimated position, it is determined to be the image of another light-emitting equipment regardless of being rectangular shape or not. The distance between vehicles is calculated from the interval between two emitting equipments. Even if noise such as a reflection light is superposed on one emitting equipment image, since the position can be estimated from the other image, the distance between vehicles can surely be measured.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inter-vehicle distance detecting device,
In particular, the present invention relates to a device that detects an inter-vehicle distance from a preceding vehicle image obtained by a camera.

[0002]

2. Description of the Related Art In recent years, various devices have been proposed for traveling following a preceding vehicle. Needless to say, the center of control during follow-up traveling is how to accurately and quickly detect the inter-vehicle distance from the preceding vehicle.

For example, in Japanese Patent Laid-Open No. 2-232515, images of two characteristic points (for example, a high-intensity portion such as a tail lamp) on the rear portion of a preceding vehicle are acquired by an image sensor, and
There is disclosed a measuring device that detects a vehicle-to-vehicle distance from a preceding vehicle from an interval between points.

[0004]

However, in order to measure the vehicle-to-vehicle distance by this method, it is premised that the two image-to-image distances are accurately measured, and one of the two images has a brightness due to dirt or the like. However, when the noise is reduced or noise such as reflected light of the sun is included, there is a problem that only one image can be extracted and the inter-vehicle distance cannot be measured.

The present invention has been made in view of the above problems of the prior art, and its object is to provide a case where one of two images does not accurately reflect an actual high brightness portion due to noise or dirt. However, it is an object of the present invention to provide a device that can reliably detect an inter-vehicle distance.

[0006]

In order to achieve the above object, a first aspect of the present invention captures two high-intensity shapes mounted on a rear portion of a preceding vehicle at a predetermined distance from each other, and first obtains an image from the obtained images. In an inter-vehicle distance detecting device for detecting an inter-vehicle distance to a vehicle, an image pickup means for photographing a rear part of a preceding vehicle, an extracting means for extracting one high-intensity shape from the obtained image, and a size of the extracted high-intensity shape It is characterized in that it has an estimating means for estimating an image position of another high-intensity shape, and a calculating means for calculating a high-intensity region at the calculated position and an inter-vehicle distance from the extracted high-intensity shape to a preceding vehicle. .

Since the mounting intervals of the two high-intensity shapes are constant, when one of the two high-intensity shapes is detected, the position of the other high-intensity shape is simply proportional to the size thereof. It can be estimated from the relationship. Therefore, if a high-luminance region exists at the estimated position, it can be considered that the region is a high-luminance image. It should be noted that the detection of one high-intensity shape is determined by whether or not it matches a shape that is known in advance. For example, when the high-intensity shape is a quadrangle, it can be detected by whether or not the high-intensity image is a quadrangle. .

Further, in order to achieve the above object, in the second invention, in the first invention, the high-intensity shape is a light emitter, and the light emission pulse from the light emitter is received to obtain the preceding vehicle data. It has a receiving means for obtaining.

There is no problem if the mounting interval of the high-intensity shape and the high-intensity shape are constant irrespective of the vehicle, but if they differ from vehicle to vehicle, it is possible to receive these data by receiving these related data from the preceding vehicle. Applicable to other vehicles.

In order to achieve the above object, a third aspect of the invention is to photograph two high-intensity shapes mounted on the rear portion of the preceding vehicle at a predetermined distance from each other, and use the obtained images to measure the distance between the preceding vehicle and the preceding vehicle. In a vehicle-to-vehicle distance detection method that detects a distance, a rear part of a preceding vehicle is photographed, a high-intensity shape is extracted from the obtained image, and other high-intensity shape positions are estimated based on the size of the extracted high-intensity shape and estimated. It is characterized in that a high-intensity region is extracted at the position, and the inter-vehicle distance to the preceding vehicle is calculated based on the extracted high-intensity shape and the interval between the high-intensity regions.

[0011]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows a block diagram of the configuration of this embodiment. Infrared light emitters 12 and 14 each having a quadrangular shape as a high-intensity shape are attached to a rear portion of the preceding vehicle 100 in a horizontal direction at a predetermined interval. This infrared light emitter 12, 14
Is supplied with data on the preceding vehicle from the data transmission circuit 10, and this data is transmitted to the following vehicle by an infrared light pulse. The data regarding the preceding vehicle includes traveling data such as the steering angle of the own vehicle, the vehicle speed, and a brake signal, as well as the shape of the infrared light emitter and the mounting interval data. In this way, by transmitting the vehicle speed and steering angle data of the preceding vehicle, the following vehicle can travel in the same manner as the preceding vehicle. The ON / OFF cycle of the pulse is 33 ms. On the other hand, the follower vehicle 200 is provided with an infrared light receiver 20 for receiving infrared light, which receives an infrared pulse and supplies it to the data receiving circuit 22. The data receiving circuit 22 demodulates the received light data and outputs it as preceding vehicle data. Further, the data from the infrared light receiver 20 is supplied to the communication / distance-measuring synchronization circuit 24 to generate a timing signal in synchronization with the timing of the light emission pulse and having a period of 66 ms. The generated timing signal is
It is supplied to the CCD camera 16 for taking a rear image of the preceding vehicle. A visible cut filter 16a is attached to the CCD camera 16 and captures a rear image of the preceding vehicle at a period of 66 ms. Therefore, the CCD camera 16 images the light emitters 12 and 14 at the light emission timing of the light emitter. The obtained image is supplied to the image processing ECU 18. The image processing ECU 18 is configured to include a binarization circuit and an image memory, calculates the distance between the two images of the infrared light emitters 12 and 14, and outputs the distance to a vehicle distance calculation ECU (not shown).

Here, since the infrared light emitters 12 and 14 are quadrangular, the image is also quadrangular in principle, and the image supplied from the CCD camera 16 to the image processing ECU 18 has two horizontal positions. This means that there is a quadrangular high-intensity part. However, if there is dirt on the infrared light emitter, reflected light of the sun, or reflected light of its own infrared light, the image will not be a quadrangle, but will change to another shape. Therefore, if the image processing ECU 18 performs processing only by the logic of extracting two quadrangular high-intensity portions, it becomes impossible to measure and the inter-vehicle distance cannot be measured.

FIG. 2 shows an example of the rear image of the preceding vehicle taken by the CCD camera 16. In the figure, 120 is an image of the infrared light emitting device 12, which has no noise and maintains a square shape. In the figure, reference numeral 140 denotes an image of the infrared light emitter 14, which is not in a quadrangular shape because the reflected light of sunlight is superposed thereon. Reference numeral 130 in the figure denotes noise due to other reflected light. Therefore, although the image of the infrared light emitter 12 can be extracted, the infrared light emitter 1 can be extracted.
It is difficult to extract the image of 4 from the noise group. Therefore, in the present embodiment, even in the case of FIG. 2, the image 140 of the infrared light emitter 14 is reliably extracted, and 120 and 140
Is calculated to calculate the inter-vehicle distance.

The processing of the image processing ECU 18 will be described in detail below with reference to the flowcharts of FIGS. 3 and 4.

First, referring to FIG. 3, the image processing ECU 18
Converts the image from the CCD camera 16 into a binary image (S101) and projects it on the XY coordinate system (S102).
The X axis is horizontal and the Y axis is vertical. Then, the light emitting area is extracted from the XY image (S103). In addition,
The light emitting area can be extracted as a set of white pixels. Then, for each of the extracted light emitting areas, it is determined whether or not the shape is a quadrangle (S104). Whether or not the shape is a quadrangle is determined by determining whether all four sides of the light emitting area are straight lines. The details will be described later. In the present embodiment, it is assumed that the number of quadrangular images including the image of the light emitter and the noise is N.

Next, the value C of the counter is reset to 0 (S105), and it is determined whether or not there is a light emitting area at the horizontal position of the largest quadrangle obtained (S106).
This horizontal position is calculated as follows. That is, the maximum extracted quadrangular light emitting region is one light emitting device 1.
Assuming that the image is image No. 2, the size of one side of the actual light emitters 12 and 14 is A, the mounting interval is L, and one of the quadrangles in the image is
The side size is a and the distance between the images of the two light emitters in the image is d.
Then

## EQU1 ## There is a relationship of d = a / AL. Since A and L are included in the preceding vehicle data output from the data receiving circuit 22, they can be used. Then, it is determined whether or not there is a light emitting area, that is, a set of white pixels, at the left and right horizontal positions at intervals of d from the maximum quadrangle. If the light emitting area exists at the position of d, it can be specified that the light emitting area is the image of the other light emitting device 14, and the largest square light emitting area is certainly the image of the light emitting device 12 (S112). ). Two
Once the two light emitter images have been extracted, using the distance n between the centers of gravity of these light emitting regions,

The inter-vehicle distance D is calculated by the following equation: D = (fL) / (nP) Note that f is the focal length of the lens, and P is the pixel pitch. In this way, the inter-vehicle distance can be reliably measured even when one of the light emitter images does not have a square shape due to noise superposition.

On the other hand, when the light emitting area does not exist at the maximum horizontal position d of the quadrangle, it can be determined that this quadrangle is not the image of the light emitter, so the counter C is incremented by 1 (S107). ), The next largest square is extracted, and the processing from S106 is repeated (S109). If the images of one set of light emitters could not be extracted even after performing the above-mentioned processing for all the extracted rectangles, it is possible that both light emitter images were affected by noise. The calculation is not performed, the previously detected inter-vehicle distance is output, and the control is ended by turning off the ignition (S111).

FIG. 4 shows a detailed flowchart of the quadrilateral recognition process performed in S104. When the light emitting area image is extracted (S201), the side counter i is reset (S202), and the uneven pixel of the side i is calculated (S20).
3). Here, the concavo-convex pixel means a pixel that is not arranged on the reference line among the pixels that are considered to form the side i.
FIG. 5 shows an example of the extracted light emitting area and its concave and convex pixels. FIG. 5 (A) shows one of the light emitting regions formed by a set of a plurality of pixels, and FIG. 5 (B) shows the uneven pixels included in one side i of the light emitting region by hatching. ing. As described above, the concavo-convex pixels are pixels that are not arranged on the reference line, and the reference line is a line formed by the largest number of pixels. The number of concave and convex pixels on the side i is three.
Then, after the number of uneven pixels on the side i is calculated,

[Equation 3] Concavity / convexity = number of concavo-convex pixels / number of circumscribing quadrangular pixels to calculate the concavity / convexity of the light emitting region (S204).
In the example of FIG. 5, the number of circumscribed quadrangular pixels is 11 × 8 = 88.
Since it is a pixel, the degree of unevenness on the side i is

## EQU4 ## 3/88 = 0.034. Then, this unevenness is compared with a predetermined threshold value (for example, 0.15) (S205). When the degree of unevenness of the side i is less than or equal to the threshold value, the side i is determined to be a straight line, the side counter i is incremented by 1, and the next side i is incremented.
The same process is repeated for +1 (S207). Then, when it is determined that all four sides are straight lines below a predetermined threshold value, this light emitting region is recognized as a quadrangle (S
208). On the other hand, if it is determined that even one of the four sides is equal to or larger than the threshold and is not a straight line, the light emitting region is recognized as a non-rectangular shape (S209).

As described above, in the present embodiment, first, the light emitting region in which all four sides are straight lines is recognized as a high-luminance shape from the obtained image, and is set as a candidate for the image of the light emitter. Then, it is determined whether or not another light emitting device image exists at the horizontal position of this light emitting device image, and if there is, another light emitting device image is detected regardless of whether the light emitting region is a square or a non-rectangle. Therefore, even if one of the light emitter images contains noise and does not accurately reflect the actual shape, its position can be detected reliably and the inter-vehicle distance can be obtained.

In the present embodiment, a quadrangular light emitter has been described as an example, but other arbitrary shapes, for example, a triangular shape or a pentagonal shape, are used.
It goes without saying that a rectangular shape can be processed in the same manner. It is also possible to make the shapes of the left and right light emitters non-targets, for example, the right light emitter 12 may be a quadrangle and the left light emitter 14 may be a triangle, and in this case, after detecting either one,
When detecting the other light emitter image, it becomes possible to know in which direction the horizontal position should be searched (when a square high-intensity shape is detected, if the left side in the horizontal direction is searched in the image). It will be good), and the processing speed can be increased. Further, as the high-luminance shape, a reflector can be applied instead of the light emitter.

Further, in the present embodiment, when one set of light emitter images cannot be detected, the previous vehicle-to-vehicle distance is output (S110), but only one detected light emitter image is approximate. It is also possible to detect the inter-vehicle distance. That is, since the estimated position d from the detected size of one side of the high-luminance shape to the image of the other light emitter is calculated, the inter-vehicle distance can be obtained according to the equation (2) using this d. In this method, for example, as shown in FIG. 6, since the preceding vehicle has changed its course, the light emitter image 1 is displayed in the front frame.
21 and 141 are shown (FIG. 6 (A)), but it is effective when only one light emitter image 121 is shown in the current frame (FIG. 6 (B)). Of course, whether to output the previous inter-vehicle distance or output the inter-vehicle distance based on only one of the light emitter images may be switched according to the traveling state of the following vehicle. For example, if the following run is relatively stable (there is no change in vehicle speed),
In a transitional situation (when there is a change in vehicle speed) in which the preceding inter-vehicle distance is output and the preceding vehicle is to be followed, it may be preferable to calculate the inter-vehicle distance from one of the light emitter images.

[0023]

As described above, according to the present invention,
2 such as light emitters and reflectors attached to the rear of the preceding vehicle
Even if noise such as reflected light is superimposed on one of the two high-brightness shapes and an image of the original shape cannot be obtained, the position can be detected reliably and the inter-vehicle distance can be calculated.

[Brief description of drawings]

FIG. 1 is a configuration block diagram of an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an example of a rear image of a preceding vehicle of the same embodiment.

FIG. 3 is a processing flowchart of the same embodiment.

FIG. 4 is a processing flowchart of the same embodiment.

FIG. 5 is an explanatory diagram of uneven pixels of the same embodiment.

FIG. 6 is an explanatory diagram showing changes in an image according to another embodiment of the present invention.

[Explanation of symbols]

10 data transmission circuit, 12 infrared light emitter (right), 14
Infrared light emitter (left), 16 CCD camera, 18 image processing ECU, 20 infrared light receiver, 22 data receiving circuit.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location // G05D 1/02 G06F 15/70 350B

Claims (3)

[Claims] 1. An inter-vehicle distance detecting device for photographing two high-intensity shapes attached to a rear portion of a preceding vehicle at a predetermined interval and detecting an inter-vehicle distance from the obtained vehicle from the obtained image. An imaging means for photographing, an extracting means for extracting one high-intensity shape from the obtained image, an estimating means for estimating an image position of another high-intensity shape from the size of the extracted high-intensity shape, and An inter-vehicle distance detection device, comprising: a high-luminance region at a position; and a computing unit that computes the inter-vehicle distance from the extracted high-luminance shape to the preceding vehicle. 2. The inter-vehicle distance detecting device according to claim 1, wherein the high-intensity shape is a light emitting device, and has a receiving unit that receives a light emission pulse from the light emitting device to obtain preceding vehicle data. Inter-vehicle distance detection device. 3. An inter-vehicle distance detecting method for photographing two high-intensity shapes attached to a rear portion of a preceding vehicle at a predetermined distance and detecting an inter-vehicle distance from the preceding vehicle from the obtained images. A high-brightness shape is extracted from the captured image, other high-brightness shape positions are estimated based on the size of the extracted high-brightness shape, a high-brightness area is extracted at the estimated position, and the extracted high-brightness area is extracted. An inter-vehicle distance detecting method for calculating an inter-vehicle distance to a preceding vehicle based on a shape and an interval between high brightness areas.