JPH11139225A - Tunnel detector device and vehicle control device using it - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a tunnel detector device with no tunnel picture of perfect shape required. SOLUTION: A forward road is photographed by a camera 1. In a white line detection part 2, a white line is detected from a road picture by function approximation. In a low concentration region detection part 4, the road picture is processed by a threshold value, and a binary picture is formed. A picture element of low concentration is extracted from the binary picture, to be classified in a plurality of regions by continuity judgment. In a notice region arithmetic part 5, a low concentration region is extracted in a direction along the white line, and in a position of start point and the uppermost point in the case of extracting a low concentration region, a distance of region to a vehicle and a virtual height are calculated. In a tunnel detection part 6, by decision whether a moving amount of the vehicle is conformed or not to a distance change amount calculated by a detected distance change of a region of low concentration, an inlet of a tunnel is detected. In a control part 8, when the tunnel inlet is detected, for preparing for running in the tunnel, for instance, control switching air intake of a vehicle air conditioning to internal air circulation is performed, and vehicle operation in the case of advancing into the tunnel is omitted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a tunnel detecting device for detecting a tunnel on a traveling road by image processing and a vehicle control device using the same.

[0002]

2. Description of the Related Art Conventionally, as a tunnel detecting device and a vehicle control device using the same, for example, Japanese Patent Laid-Open No.
Is disclosed in Japanese Unexamined Patent Application Publication No. 2000-205,878. Here, a one-dimensional line sensor is used to observe the road ahead. A black area below the set density is detected from the output of the one-dimensional line sensor, and the length is determined to detect a tunnel. In Japanese Patent Application Laid-Open No. 60-240545, a two-dimensional image sensor is used to observe a road ahead. The road image is subjected to image processing to detect a black area having a density equal to or less than a set density, and determine whether or not the black area is a tunnel entrance by comparing it with a reference pattern to detect a tunnel.

[0003]

However, although the width of the tunnel or the shape of the tunnel is various, if the width is limited to the traveling path, the width becomes a size corresponding to the traveling path.
The shape also covers the width of the traveling path. Therefore, in order to recognize these, in the detection based on the above-mentioned conventional pattern, the tunnel needs to be observed in perfect form.
Therefore, in the observation on the traveling road, there is a problem that the tunnel cannot be detected if the visibility is obstructed by the presence of the preceding vehicle and the tunnel cannot be completely observed.

In addition, when the tunnel is short and the scenery after the exit is visible at the entrance of the tunnel, there is a problem that the width and shape of the black area are different from the usual ones, and the pattern cannot be detected. The present invention has been made in view of the above problems,
An object of the present invention is to provide a tunnel detection device that does not require a complete tunnel image and can detect a tunnel from a vehicle even when following a preceding vehicle, and a vehicle control device using the tunnel detection device.

[0005]

According to the present invention, there is provided a tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels, wherein the image capturing means captures an image of a traveling direction of the vehicle. A traveling path detecting means for detecting a direction in which the traveling path continues, and a set density area detecting means for detecting an area having a predetermined density from an image captured by the imaging means;
Attention area extraction means for searching for a direction following the traveling path and extracting an area of a predetermined density to be an attention area indicating the entrance or exit of a tunnel, attention area calculation means for calculating an image position of the attention area, If the distance change calculated by the change in the image position of the attention area matches the moving amount of the vehicle and the apparent height of the attention area exceeds a predetermined value, the attention area is determined to be the entrance or exit of the tunnel. And a tunnel detecting means.

According to a second aspect of the present invention, there is provided a tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels.
Travel path detecting means for detecting a direction in which the travel path continues,
Setting density area detecting means for detecting a low-density area from the image captured by the imaging means; and searching for a direction following the traveling road to extract the low-density area, and as a target area indicating a tunnel entrance. An area extracting means, an attention area calculating means for calculating an image position of the attention area, and a distance change calculated by a change in the image position of the attention area coincides with a moving amount of the vehicle, and an apparent height is a predetermined value. When the distance exceeds the threshold, a tunnel entrance detecting means for determining the region of interest as a tunnel entrance is provided.

According to a third aspect of the present invention, there is provided a tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels;
Travel path detecting means for detecting a direction in which the travel path continues,
Setting density area detecting means for detecting a high density area from an image picked up by the imaging means; and searching for a direction following a traveling road to extract the high density area, and setting it as a target area indicating a tunnel exit. Region extraction means, attention area calculation means for calculating the image position of the attention area, and the distance change calculated by the image position change of the attention area coincides with the moving amount of the vehicle, and the apparent height of the attention area If the value exceeds a predetermined value, a tunnel exit detecting means for judging the attention area as a tunnel exit is provided.

According to a fourth aspect of the present invention, there is provided a tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels;
Setting density area detecting means for detecting an area of a predetermined density from the image picked up by the imaging means; area calculating means for calculating an area of the area detected by the setting density area detecting means on the image; As a result, the calculated value of the area calculating means increases, and the calculated value of the area becomes equal to or larger than the set value.

[0009] The travel path detecting means can process the captured image to detect a white line, and detect the direction in which the travel path continues by function approximation.

According to a sixth aspect of the present invention, a tunnel traveling control means for controlling a vehicle to a state suitable for traveling in a tunnel when the tunnel entrance detecting means determines a tunnel entrance is connected to the tunnel detecting device. It was done.

The invention according to claim 7 is characterized in that the tunnel detecting device includes a tunnel travel canceling means for performing control for returning the vehicle from the tunnel traveling state to the normal state when the tunnel exit detecting means determines the exit of the tunnel. Connected.

[0012]

According to the first aspect of the present invention, the set density area detecting means detects an area having a predetermined density from the front road image captured by the image capturing means. Since the tunnel is imaged under different lighting conditions than the surrounding environment, the entrance or exit of the tunnel appears with different densities on the image. Therefore, by detecting them as a predetermined concentration, an entrance or exit region indicating the presence of a tunnel is detected.

Since the entrance or the exit of the tunnel exists on the traveling path of the vehicle, the traveling path detecting means detects the direction in which the traveling path continues, and the attention area extracting means searches for the direction in which the traveling path continues, and searches for the predetermined direction. By extracting the density area,
Other regions of the same density are excluded, and only the region of interest indicating the entrance or exit of the tunnel is extracted.

The tunnel detecting means can determine whether or not the attention area is a fixed object such as a tunnel by comparing the moving amount of the vehicle with the distance change amount calculated from the position of the attention area on the image. If the area of interest is the entrance or exit of the tunnel, the apparent height is also the height of the tunnel. Therefore, the attention area extracted by the determination based on the predetermined value can be determined as the entrance or the exit of the tunnel. As a result, a tunnel is detected. in this way,
As long as data indicating the height and distance of the tunnel can be obtained, a complete tunnel image is not required, and the tunnel can be detected from the traveling road even if the view of the preceding vehicle is obstructed.

According to the second aspect of the present invention, the detection area is a low-concentration area, and the entrance of the tunnel is detected as the attention area.

According to the third aspect of the present invention, the detection area is a high-concentration area, and the exit of the tunnel is detected as the attention area. Therefore, the invention is suitable for tunnel detection in the daytime.

According to the fourth aspect of the invention, the set density area detecting means detects an area of a predetermined density from the tunnel image picked up by the image pick-up means. Since the illumination inside and outside the tunnel is different, the exit image of the tunnel is captured with a different density from the surroundings. As the vehicle travels, the exit of the tunnel is imaged gradually larger, so that the tunnel detecting means can detect the exit of the tunnel by increasing the area of the detected region of the predetermined density. As a result, a tunnel is detected. As described above, a tunnel can be detected only by detecting a region having a predetermined density and calculating an area.

When the traveling road detecting means detects the white line by processing the captured image and detects the direction in which the traveling road continues by function approximation, the location where the tunnel exists can be specified accurately, and the reliability of the tunnel detection can be improved. The performance is improved.

According to the sixth aspect of the invention, when the entrance of the tunnel is detected, the tunnel traveling control means controls the vehicle to a state suitable for traveling in the tunnel. The driving load is reduced and safe driving is achieved.

According to the seventh aspect of the present invention, when the exit of the tunnel is detected, the tunnel traveling control means controls the vehicle to a state suitable for traveling outside the tunnel, so that the vehicle is set to travel inside the tunnel. The function can be automatically released, and the operation for releasing the function can be omitted. In accordance with claim 6, the vehicle can be driven without being conscious of the tunnel.

[0021]

Embodiments of the present invention will be described below with reference to examples. FIG. 1 is a block diagram showing the configuration of the first embodiment. The camera 1 as an imaging means is mounted on the vehicle slightly downward toward the front, so that the camera 1 can capture the traveling direction of the vehicle including the traveling path. The camera 1 is provided with an image memory for storing image data.
The image data is stored here, and the data is updated by inputting new image data.

The camera 1 is connected to a white line detecting section 2 and a low density area detecting section 4. The white line detection unit 2 detects white lines on the left and right sides of the image by image processing and detects them by function approximation in order to detect the own vehicle traveling road. The farthest point of the white line detected in the image processing is normalized by a function approximation formula and stored in a memory together with the function approximation formula.

The low-density area detector 4 binarizes the image data with a predetermined threshold to generate a binarized image.
A low-density pixel equal to or less than a threshold value is extracted from the binary image, and is classified into a plurality of regions by determining continuity. The position data of each classified area is stored in the memory. Since the tunnel is darker than the surrounding environment in the daytime, the entrance area is imaged with low density,
It will be stored in memory.

The attention area extraction unit 3 which inputs the detection result of the white line detection unit 2 and the position data of the area extracted by the low density area detection unit 4 converts the white line displayed by function approximation from the farthest point position of the white line. A search is made for a low-density area along the distance. As a result, the low-concentration area outside the travel path is excluded, and the attention area indicating the entrance of the tunnel is extracted. Then, in order to confirm whether or not the attention area is the entrance of the tunnel, extraction is performed by searching for the highest point of the low density area. By extracting the highest point, the height of the tunnel can be calculated, and even if the tunnel overlaps with the preceding vehicle, the tunnel can be determined based on the height.

The attention area calculation unit 5 determines the apparent height of the attention area and the distance to the vehicle based on the imaging conditions of the camera 1, that is, the focal length of the camera, the imaging angle, the mounting height of the camera, etc., based on the position data of the attention area. Is calculated. The tunnel detector 6 calculates the travel distance of the vehicle based on the traveling speed of the vehicle obtained by the vehicle speed sensor 7 and the update cycle time of the image. If the movement distance matches the change in distance calculated by the change in the position of the low-density region, the region of interest is determined to be the entrance of the tunnel, and the tunnel is detected after confirmation based on the apparent height. If a tunnel is detected here, a flag is set to indicate that a tunnel has been detected.

The control unit 8 is connected to the tunnel detection unit 6. When the flag is set, the control unit 8 turns on the headlight in preparation for traveling in the tunnel, checks the air intake state of the air conditioner, and performs control to switch to internal air circulation if external air is being introduced. Further, for example, control such as how to apply an engine brake, a gain of a throttle, an operation interval of a wiper, closing of a sunroof, and the like are performed according to a vehicle, and the vehicle is controlled to a state suitable for traveling in a tunnel. For a camera without an automatic exposure function, the exposure is adjusted to a value suitable for imaging in a tunnel.

Next, the flow of processing of the above configuration will be described with reference to the flowchart of FIG. First, in step 101, the road ahead is imaged by the camera 1, and the image data of the imaged is stored in the image memory. Thereby, for example, an original image as shown in FIG. 3 is obtained. In the image of FIG. 3, a tunnel 13 is shown on a traveling path where white lines 11 and 12 continue to the front, and a part of the tunnel is
Has disappeared.

In step 102, the white line detection unit 2 receives image data from the image memory, performs image processing to detect white lines on the left and right sides of the travel path, and recognizes them by function approximation.
Then, the position of the white line at the farthest point detected in the image processing is normalized by a function approximation formula, and stored in a memory together with the function approximation formula. Thus, in the original image of FIG. 3, white lines are displayed as curves 21 and 22 represented by function approximation formulas, as shown in FIG. Points a and b are the farthest points of the normalized left and right white lines. The details of the white line detection will be described later.

In step 103, the low density area detecting section 4
To process the original image with a predetermined threshold value, and create a binary image having a density value 1 when the value is equal to or less than the threshold value and a density value O when the other value is less than the threshold value. As a result, a binarized image is formed from the image of FIG. 3 in which the dark tunnel portion 13 has the low density region 13 with the density value 1 as shown in FIG. In this image, part of the tunnel is missing due to the preceding vehicle.

In step 104, the attention area extraction unit 3 searches the low density area along the direction in which the left and right white lines extend from the farthest point to extract the low density area indicating the tunnel entrance as shown in FIG. , And extract the highest point position for calculating the apparent height. FIG. 6 shows the state of the processing. That is, first, a low density area is searched from the farthest point a in the direction indicated by the function approximation formula of the white line.

When the low density area (pixel density 1) is detected, the coordinates are used as the starting point coordinates and the density value is rewritten as 2.
From the starting point coordinates, the X coordinate is searched in both the positive and negative directions, and if there is an adjacent low density area (pixel density 1), the density value is rewritten as 2. Next, by adding one Y coordinate, a low-density area (pixel density 1) adjacent to the pixel having the density value 2 is searched above the image (in the far direction). If a pixel satisfying the above condition is found, the density value is rewritten to 2. This process is repeated until a low-density pixel cannot be detected. If a pixel satisfying the above conditions cannot be detected any more, the Y coordinate is reduced by 1 to be the upper limit of the low density area connected to the left white line (curve 21).

Next, the right white line is searched for a region connected to the right white line (curve 22) from the farthest point b by the same operation as described above, and if a low density region is detected, the density value is rewritten as 3. . If a pixel having a density value of 2 is detected during the search, it is determined that the left and right white lines are connected to the same low density area, and the upper limit value of the left white line area is stored as the highest point of the low density area.
As a result, the point c is detected in FIG.

In step 105, the attention area calculation unit 5 calculates the distance D to the vehicle and the apparent height H using the low density area extracted from the white line as the attention area indicating the entrance of the tunnel. The starting point coordinates of the attention area are at the bottom of the area, and the forward distance to the vehicle is calculated from the camera mounting position, angle, and parameters of the lens system. Here, first, a conversion formula for the forward distance and the Y coordinate is obtained in advance according to the imaging conditions of the camera, and the forward distance is calculated by substituting the starting point coordinates into the conversion formula.

Although the height H cannot be directly detected by a single-lens camera, it can be assumed that a region of interest stands perpendicularly to the origin coordinate position from the characteristics of the tunnel entrance. Therefore, from the camera mounting position, the angle, and the parameters of the lens system, a conversion formula is obtained in advance in the same way as the forward distance is obtained, and the apparent coordinates of the attention area are obtained by substituting the starting point coordinates and the top point coordinates into the conversion formulas. Calculate the height.

FIG. 7 is an explanatory diagram for obtaining a conversion formula between the forward distance and the position on the image. (A) shows the vertical direction, and (b) shows the geometric relationship between the imaging space and the planar image projected in the horizontal direction. Here, the camera has a lens having a focal length f and an imaging CCD arranged at the focal length f. The camera is attached to the vehicle so as to capture an image downward at an angle θ with respect to the horizontal. The mounting height of the camera is h
It has become. The CCD has IX pixels in the X direction and I pixels in the Y direction.
It is composed of Y pixels. The screen size of the CCD is CH (mm) in the X direction and CW (mm) in the y direction.

First, in (a), equation (1) holds for an angle dθ formed by the imaging direction of the camera. dθ = atan [(CH / IY) · (dy / f)] (1) where dy is the image position in the Y-axis direction with respect to the center of the screen. The relationship between dθ and the forward distance D is shown by equation (2). D = h / tan (θ−dθ) (2) CH, IY, f, h, and θ are known, and a conversion formula between the pixel position and the forward distance is obtained by Expressions (1) and (2). Thus, the forward distance D can be calculated by substituting the Y coordinate of the starting point of the low density area into the equation (1).

Although the pixel position dx in the X direction does not participate in the calculation of the forward distance, the width of the object at a predetermined distance ahead can be calculated based on the size thereof, so that the threshold for judging the size of the area in the next embodiment will be described. Useful for making decisions. The point is the same as above, and in FIG. 7B, Expression (3) holds for an angle dφ formed with the direction indicating the forward distance. dφ = atan [(CW / IX) · (dx / F)] (3) where F is F = [[(CH / IY) · dy] ² + f ² ]
It is ^1/2 . The width E corresponding to the pixel position dx is given by Expression (4)
Indicated by E = (h ² + D ² ) ^1/2 · tan (dφ) (4) With this, the forward distance D and the width are obtained at the image positions dx and dy, and the area can be calculated. The conversion formula for the distance in the Y-axis direction and the apparent height H can also be calculated from (a) as described above.

In step 106, the traveling speed of the host vehicle is measured by the vehicle speed sensor 7. In step 107, the tunnel detection unit 6 estimates how much the vehicle has moved during that time using the time required for the image update cycle and the traveling speed of the vehicle, and sets the estimated movement amount to ΔL.

In step 108, the amount of movement of the vehicle and the amount of change in the distance of the region of interest are collated, and if they match, the apparent height H of the region of interest is compared with a predetermined threshold to detect a tunnel. The details will be described later. In step 109, it is checked whether or not a tunnel has been detected. If the tunnel has not been detected, the process returns to step 101. If detected, a flag is set and the process proceeds to step 110. In step 110, the control unit 8 performs control to switch the air circulation path of the air conditioner to the internal air circulation, for example, in preparation for traveling in a tunnel by setting the flag.

FIG. 8 is a flowchart showing details of the tunnel detection in step 108. First, in step 201, the moving amount ΔL of the vehicle is compared with the threshold value th1 to determine the amount of the moving amount. The process proceeds to step 202 so as to evaluate the distance D.
In step 202, the apparent height H of the attention area is compared with a threshold th2 to determine the size.
If it is larger than th2, the process proceeds to step 203 so as to make the next determination.

In step 203, the difference between the distance D and the previously calculated distance is calculated to obtain a distance change amount ΔD, which is compared with the vehicle movement amount ΔL. If the values are about the same,
Proceed to step 205.

In step 205, it is checked whether or not the same result continues three times (th3) even if the image changes as a result of the above determination. If the same result is obtained three times, it is determined in step 207 that the attention area is the tunnel entrance. If the check condition in step 205 is not satisfied, it is determined in step 208 that there is no tunnel. Step 206 is a process for counting the number of measurements, and step 204 is a process for clearing the count.

Next, the details of the white line detection in step 102 will be described with reference to the flowchart of FIG. First, in step 300, image data is input from the camera 1. In step 301, a vertical edge image as shown in FIG. 10, for example, is created by multiplying the entire image by a SOBEL operator for extracting a vertical edge component. In this image, the edge of the white line is displayed. Then, as the density changes from left to right, the edge characteristics change as a positive edge, a negative edge, a positive edge, and a negative edge. 1 for negative and positive edges
Two white lines are detected.

In step 302, window position information is fetched. Initially, for example, a window position corresponding to a state in which the vehicle is traveling straight in the center of the lane assuming a straight road may be stored in the memory as data. From the second time, a window is set with the window position information calculated based on the processing result of the previous screen. As shown in FIG. 11, this position information is composed of the white line position Wp detected last time and the inclination γ of the white line recognized by approximation. The window W is a trapezoidal window centered on Wp and having the upper side of the width Wl and a predetermined height s. Y both sides of window from top
The angles are γ + α and γ-α outward from the axis. α is a fixed value set in advance according to the mounting state of the camera. The window is a small section, and a white line having a different shape can be regarded as a straight line here.

The straight line detection at step 303 will be described with reference to the flowchart of FIG. First, in step 402, a window W is searched and edge points are totaled to determine a straight line. For this, as shown in FIG. 13, the initial position T (xt, yt) of the straight line tip is changed to the window position Wp (xwp, y).
wp) and the upper side width Wl. ywp is a set value indicating the position of the window W. Tip initial position T (xt,
yt) is xt = xwp + W1 / 2/2 yt = ywp.

In step 403, a straight line rear end initial position E (xe, ye) is calculated from the window height s. xe = xt-s · tan (γ-α) ye = yt-s

In step 404, T (xt, y
A straight line f passing through t) and E (xe, ye) is calculated. The straight line f is y-ye = (yt-ye) / (xt-xe) · (xx
e)

In step 405, the number of edge points on the edge image through which the straight line f passes is measured. At this time, the numbers of both the positive edge and the negative edge are separately measured. The number of positive and negative edges measured together with the leading end position xt and the trailing end position xe is stored in a memory. In step 406, the x coordinate of the rear end E of the straight line is moved so that xe = xe-1. In step 407, xe is xt-s · ta
It is determined whether or not the rear end has scanned the window W by determining whether or not it is greater than n (γ + α). If the window has not been scanned, the process returns to step 404. If the window has been scanned, the process proceeds to step 408.

In step 408, xt-1 is substituted for the x coordinate xt of the straight tip T to move the tip position. In step 409, it is determined whether or not xt of the straight line tip is equal to or less than xwp-Wl / 2, and it is determined whether the tip has scanned the window. If the scanning has not been performed, the process returns to step 404 and the above processing is continued. If the window has been scanned, the process proceeds to step 410.

In step 410, the leading edge position and the trailing edge position of the maximum number of positive edges are extracted from the stored leading edge positions and the number of positive edges. In step 411, the number of extracted positive edges is compared with a threshold value. If the number is greater than the threshold value, the process proceeds to step 412. In step 412, the extracted front end position and rear end position are connected by a straight line, and an approximate line of the positive edge is determined.

In step 413, the leading edge position and the trailing edge position of the maximum negative edge number are extracted from the number of negative edges. In step 414, the number of negative edges is compared with a preset threshold value.
If the value is equal to or smaller than the threshold value, the detection is terminated and the process proceeds to step 304. In step 415, the extracted front end position and rear end position are connected by a straight line, and an approximate line of the negative edge is determined.

Thereafter, returning to the flow chart of FIG. 9, it is checked in step 304 whether or not the approximate lines 31 and 32 of the positive and negative edges have been detected from the window W as shown in FIG. If there is a white line, the process proceeds to step 305;

In step 305, an approximate line of a white line passing between the approximate lines of the positive and negative edges is calculated. The white line is a straight line, and the parameters for determining it include Wp for determining the position and γ for determining the slope. Wp (xw
The x coordinate xwp of (p, ywp) is calculated by (xt + xe) / 2. ywp is the setting position of the window.
In step 306, the window setting position of white line detection in the next image is calculated using Wp and γ as window position information. As a result, the window fluctuates in accordance with the position fluctuation of the white line, so that even if the position of the white line moves, the white line can be detected without leaking. This completes the white line detection process.

Although the detection of a straight line has been described in one window, when a white line is approximated by a curve, a plurality of windows as described above are set for one white line, and the leading and trailing end positions of each window are determined. By the function approximation, for example, an approximation curve as shown in FIG. 4 is obtained. White line detector 2
Constitutes a traveling path detecting means. Low density area detector 4
Constitutes a set density area detecting means. The attention area extraction unit 3 constitutes attention area extraction means. The attention area calculation unit 5 constitutes attention area calculation means. The control section 8 constitutes tunnel traveling control means.

The present embodiment is configured as described above, and detects a white line indicating a traveling path, and detects a region of interest indicating a tunnel entrance from a direction in which the white line continues. Then, whether the movement amount of the vehicle and the change in the distance from the attention area to the vehicle match and the apparent height of the attention area is determined to detect the tunnel. In addition, even if the tunnel is not visible in front of the vehicle due to a curve or the like, there is an effect that the tunnel on the road can be detected from the white line of the road.

Next, a second embodiment will be described. In this embodiment, the exit of the tunnel is to be detected. FIG.
FIG. 2 is a diagram showing a configuration of an embodiment. The camera 1 is mounted in front of the vehicle so as to capture an image of the traveling direction of the vehicle as in the first embodiment. The low-density area detection unit 4 generates a binarized image by binarizing the image captured by the camera 1 with a predetermined threshold. The area calculation unit 9 calculates the area by measuring the number of pixels in an area other than the low density area.

The tunnel detector 10 detects a tunnel based on a change in the area calculated by the area calculator 9 as the vehicle travels. When a vehicle travels in a tunnel, images are taken at a low density except for the exit area of the tunnel. The more the vehicle approaches the exit of the tunnel, the larger the exit is observed,
The exit of the tunnel can be detected from the change in the area of that part. When the exit is detected, the control unit 8 performs control to release the function set for traveling in the tunnel. For example, the ventilation circulation path of the air conditioner is switched to take in outside air.

Next, the operation flow of the above configuration will be described with reference to the flowchart of FIG. First, in step 501, the road ahead is imaged by the camera 1, and the imaged image data is stored in the image memory. Thereby, for example, an image as shown in FIG. 18 is stored. The image shown in FIG. 17 is an image taken in the tunnel, and has a low density as a whole, and only the exit 23 of the tunnel is displayed with a high density on a traveling path where the white lines 11 and 12 continue to the front. Part of the exit is invisible due to the preceding vehicle 14.

In step 502, the low-density area detection unit 4 processes the image with a threshold value, and creates a binary image having a density value 1 when the value is greater than the threshold value and a density value O when the value is less than the threshold value. As a result, a binarized image of the low density area is created from the image of FIG. 17 except for the exit portion 23 of the tunnel as shown in FIG. In this image, part of the tunnel is missing due to the preceding vehicle.

In step 503, the area calculating section 9 calculates the area R of the bright image portion by measuring pixels other than the low density area. In step 504, the vehicle speed is measured by the vehicle speed sensor 7. Step 50
In 5, the tunnel detection unit 10 estimates how much the vehicle has moved during that time using the time required for the image update cycle and the traveling speed of the vehicle, and sets the estimated movement amount to ΔL.

The detection of the tunnel exit in step 506 is shown in FIG.
This will be described with reference to the flowchart of FIG. First, in step 601, the magnitude of the movement amount is determined by comparing the movement amount ΔL of the vehicle with the threshold value th1. If it is larger than the threshold value, the process proceeds to step 602. In step 602, a difference ΔR between the calculated area R and the previous area is calculated and compared with a predetermined threshold value. Step 6 if greater than threshold
Go to 04.

In step 604, the counter is checked to determine whether the result of the above determination has continued three times (th5). As a result, assuming that the time at which the image of FIG. 17 was captured is t1, the images captured at times t2 and t3 are determined to have the tunnel exit area 23 gradually increasing as shown in FIGS. The image in FIG. 21 is a binary image at time t2, and the image in FIG. 23 is a binary image at time t3. The determination of the amount of change in the area and the size are performed based on them. If the check condition is satisfied, the process proceeds to step 606 to make the next determination.

In step 606, the calculated area R is compared with a predetermined threshold value th6 to check the size. If the area is smaller than the threshold, step 6
In 08, it is determined that there is no tunnel. If it is larger than the threshold value, the exit of the tunnel is detected in step 607. th
6 is set according to the size of the tunnel in, for example, the above equations (1), (2), (3), and (4). Step 605 is a process for counting the number of measurements, and step 603 is a process for clearing the count.

Thereafter, returning to the flowchart of FIG. 16, it is checked in step 507 whether the exit of the tunnel has been detected, and if not detected, the flow returns to step 501. If detected, the flag is set and the flow proceeds to step 508. . In step 508, the control unit 18 switches the air intake path of the air conditioner to outside air circulation, for example, to prepare for the vehicle to exit the tunnel due to the flag being set, and controls the vehicle to run outside the tunnel. The area calculating section 9 constitutes an area calculating means. The control unit 18 constitutes a tunnel travel release unit.

The present embodiment is configured as described above. By taking an image of the front running road and extracting a low-brightness area, a tunnel exit candidate is detected, and the amount of movement of the vehicle is detected as a candidate for the tunnel exit candidate area. Since the exit of the tunnel is determined by comparing the amount of change in the area with a predetermined threshold value, there is an effect that it is possible to detect even when there is a preceding vehicle or when the vehicle does not have an ideal tunnel shape. Although the above-described embodiment has described the detection in the daytime, the image density is reversed at night,
Needless to say, the above method is still effective.

[0066]

According to the first aspect of the present invention, the entrance and exit of a tunnel exhibiting a concentration different from that of the surrounding environment are detected.
Since the detection is performed from the direction following the white line, it can be easily distinguished from other objects showing the same concentration, the tunnel can be detected by simple processing, and even if the tunnel is partially hidden by the preceding vehicle etc. Can be detected. This satisfies the performance required of the on-vehicle device, and makes it possible to specify the entrance or exit of the tunnel and automatically perform vehicle operations for entering or exiting the tunnel.

According to the second aspect of the present invention, since the detection area is a low density area and the entrance of the tunnel is detected as the attention area, it is suitable for tunnel detection in the daytime.

According to the third aspect of the present invention, since the detection area is set as the high density area and the exit of the tunnel is detected as the attention area, it is suitable for the tunnel detection in the daytime.

According to the fourth aspect of the invention, the set density area detecting means detects an area having a predetermined density from the tunnel image picked up by the image pick-up means. Since the illumination inside and outside the tunnel is different, the exit image of the tunnel is captured with a different density from the surroundings. As the vehicle travels, the exit of the tunnel is imaged gradually larger, so that the tunnel detecting means can detect the exit of the tunnel by increasing the area of the detected region of the predetermined density. As a result, a tunnel is detected. As described above, the tunnel can be detected only by detecting the area of the predetermined density and calculating the area, and an effect that does not require a complicated calculation can be obtained.

When the traveling path detecting means detects the white line by processing the captured image and detects the direction in which the traveling path continues by function approximation, the location where the tunnel exists can be specified accurately, and the reliability of the tunnel detection can be improved. The performance is improved.

According to the sixth aspect of the present invention, when the entrance of the tunnel is detected, the tunnel traveling control means controls the vehicle to a state suitable for traveling in the tunnel. The driving load is reduced and safe driving is achieved.

According to the seventh aspect of the present invention, when the exit of the tunnel is detected, the tunnel traveling control means controls the vehicle to a state suitable for traveling outside the tunnel, so that the vehicle is set to travel inside the tunnel. The function can be automatically released, and the operation for releasing the function can be omitted. In accordance with claim 6, the vehicle can be driven without being conscious of the tunnel.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a first embodiment.

FIG. 2 is a flowchart illustrating a flow of processing of the entire embodiment.

FIG. 3 is an original image obtained by capturing a tunnel.

FIG. 4 is a diagram showing a state where a white line is detected.

FIG. 5 is a binary image subjected to a binarization process.

FIG. 6 is an explanatory diagram for searching for a low density area and detecting a starting point and an uppermost point.

FIG. 7 is an explanatory diagram for obtaining a conversion formula between an image position and a forward distance.

FIG. 8 is a flowchart for determining a tunnel.

FIG. 9 is a flowchart for detecting a white line.

FIG. 10 is a vertical edge diagram.

FIG. 11 illustrates window setting.

FIG. 12 is a flowchart for detecting a straight line.

FIG. 13 is an explanatory diagram for detecting an edge.

FIG. 14 is an image in which an edge approximate line is detected.

FIG. 15 is a block diagram showing the configuration of the second embodiment.

FIG. 16 is a diagram showing a flow of processing of the entire embodiment.

FIG. 17 is a captured image in a tunnel.

FIG. 18 is a binarized image.

FIG. 19 is a flowchart for determining a tunnel.

FIG. 20 is a captured image after a predetermined time has elapsed.

FIG. 21 is a binary image.

FIG. 22 is a binarized image after a predetermined time has elapsed.

FIG. 23 is a binary image.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Camera 2 White line detection part 3 Attention area extraction part 4 Low density area detection part 5 Attention area calculation part 6 Tunnel detection part 7 Vehicle measurement sensor 8, 18 Control part 9 Area calculation part 10 Tunnel detection part 11 Left white line 12 Right white line 13 Low-concentration tunnel image 14 Leading vehicle 21, 22 Approximate curve 31 Positive edge approximation line 32 Negative edge approximation line a, b Farthest point c Top point D Distance E Width h Camera mounting height θ Camera imaging angle with respect to horizontal plane W window s Window height Wl Window upper edge width γ Straight line inclination angle Wp Window upper edge center point

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code FI G08G 1/16 G06F 15/62 380

Claims (8)

[Claims] 1. A tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels, wherein the image capturing means captures an image of a traveling direction of the vehicle, and a traveling path detecting means detects a direction in which the traveling path continues. A set density area detecting means for detecting a predetermined density area from the image picked up by the image pick-up means; and an area of interest indicating an entrance or an exit of a tunnel by searching a direction following the traveling road and extracting a predetermined density area. An attention area extracting means, an attention area calculation means for calculating an image position of the attention area, and a distance change calculated by a change in the image position of the attention area coincides with the moving amount of the vehicle, and an appearance of the attention area A tunnel detecting means for determining the region of interest as an entrance or an exit of a tunnel when the height of the tunnel exceeds a predetermined value. 2. A tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels, wherein the image capturing means captures an image of a traveling direction of the vehicle, and a traveling path detecting means detects a direction in which the traveling path continues. A set density area detecting means for detecting a low density area from a captured image of the image capturing means, and a direction of continuity of a traveling path is searched to extract the low density area, and an attention area indicating an entrance of a tunnel. An attention area extracting means, an attention area calculation means for calculating an image position of the attention area, and a distance change calculated by the image position change of the attention area coincides with a moving amount of the vehicle, and an apparent height is reduced. A tunnel detection device, comprising: a tunnel entrance detecting means for determining the region of interest as a tunnel entrance when a predetermined value is exceeded. 3. A tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels, wherein the image capturing means captures an image of a traveling direction of the vehicle, and a traveling path detecting means detects a direction in which the traveling path continues. A set density area detecting means for detecting a high density area from a captured image of the imaging means, and a high density area extracted by searching for a direction in which a traveling road continues, and an attention area indicating a tunnel exit. Region of interest extraction means, region of interest calculation means for calculating the image position of the region of interest, and the distance change calculated by the image position change of the region of interest coincides with the amount of movement of the vehicle, and the apparent region of the region of interest When the height exceeds a predetermined value, a tunnel exit detecting means for judging the attention area as a tunnel exit is provided. 4. A tunnel detecting device provided in a vehicle for detecting a tunnel as the vehicle travels, wherein the image capturing unit captures an image of a traveling direction of the vehicle, and a region having a predetermined density is detected from an image captured by the image capturing unit. A set density area detecting means, an area calculating means for calculating an area of the area detected by the set density area detecting means on the image, and the calculated value of the area calculating means increases due to the movement of the vehicle, and the area And a tunnel exit detecting means for judging the exit of the tunnel based on the calculated value of becomes not less than a set value. 5. The travel path detection means performs image processing on the captured image to detect a white line, and detects a direction in which the travel path continues by function approximation.
5. The tunnel detecting device according to 2, 3, or 4. 6. A tunnel driving control means for controlling a vehicle to be in a state suitable for traveling in a tunnel when the tunnel entrance detecting means determines the entrance of the tunnel, to the tunnel detecting device. The vehicle control device according to claim 2. 7. A tunnel travel canceling means for controlling a vehicle to return from a tunnel traveling state to a normal state when the tunnel exit detecting means determines a tunnel exit is connected to the tunnel detecting device. The vehicle control device according to claim 3 or 4, wherein 8. The tunnel traveling control means controls an air circulation path of an air conditioner, an effect of an engine brake, a throttle gain, a wiper, a headlight, a window, a sunroof or an imaging means, and controls an air circulation path of the air conditioner to an inside air. Switch to circulation, increase engine braking force, reduce throttle gain, extend the interval between wipers, turn on headlights, close windows and sunroof, or increase exposure of imaging means 7. The vehicle control device according to claim 6, wherein: