JPH02118705A - Picture processor of mobile vehicle - Google Patents

Abstract

PURPOSE:To attain a picture processing even in travel conditions when a luminous quantity lacks so as to detect the ends of a road by deciding the ends of the travel road based on information on an area upper than an infinite line when the external luminous quantity is less than a prescribed value. CONSTITUTION:When the external luminous quantity is less than the prescribed value, the ends of the travel road is decided based on information on the area upper than the infinite line. Namely, the road lights and the like of the travel road, whose luminous quantity lacks, are continuously arranged, picture data of the road lights photographed on the upper part of the screen are processed and the ends of the travel road is detected. When a vehicle travels the center of a tunnel, for example, the position of a virtual center line 35 is obtained from the projection of a center point Q between sodium lights 30a and 30b for the road surface. Consequently, travel control can be executed by setting the virtual center line 35 to be an objective point. Thus, picture data of the road lights and the like photographed at the upper part of the screen of an image-pick up picture are processed so as to detect the ends of the travel road even if the actual end of the path cannot be recognized due to the lack of the luminous quantity.

Description

[Detailed description of the invention] (Industrial application field) The present invention uses an image of a driving route to detect, for example, a road edge.

The present invention relates to an image processing device for a moving vehicle for detecting road edges such as white lines.

(Prior Art) In a mobile vehicle that recognizes the outside world and performs autonomous driving control, a tangled edge or the like is generally detected and running control is performed along this tangled edge. For example, JP-A-61-100
No. 898 sequentially specifies passing singular points, sets a planned route to a destination, and guides the vehicle along the route. Also, JP-A No. 61-240307
As shown in No. 1, an image is captured and the travel of the vehicle is controlled based on the line at infinity in the image.

(Problem to be Solved by the Invention) By the way, especially when capturing an image and processing the image to recognize an object, as in the above-mentioned Japanese Patent Application Laid-Open No. 61-240307, When you are not at the level,
The image becomes unclear. This is because the tied end or center line itself usually does not emit light, and at night, an image of reflected light from illumination light lacks contrast, making image processing difficult.

In addition, not only at night, but especially in tunnels, even during the day there is no sunlight and the tunnel is illuminated by lighting, so the entire image has a reddish tinge, making it difficult to detect tangled edges.

Therefore, the present invention was proposed in order to eliminate the drawbacks of the above-mentioned conventional examples, and its purpose is to enable image processing and detect tangled ends even in driving situations where the amount of light is insufficient. We propose an image processing device for moving vehicles that can perform

(Means and effects for achieving the object) The configuration of the image processing device for a moving vehicle according to the present invention for achieving the above object includes an imaging means for taking an image of the traveling path ahead in the direction of travel, and a A light amount detection means for detecting that the amount of light is below a certain value, a specifying means for specifying an area above the infinity line in the image, and an output of the light amount detection means, and when the amount of light in the outside world is less than a certain value. The vehicle is characterized by further comprising determining means for determining the end of the running road based on information of an area above the infinity line specified by the specifying means.

On driving roads where the amount of light is insufficient (for example, inside tunnels), for example, road lights etc. are installed continuously, so image data of these road lights etc. taken at the top of the screen is processed. Detects the edge of the road.

(Embodiment) An embodiment of the image processing apparatus of the present invention will be described below with reference to the accompanying drawings. In this embodiment, especially, image processing is made possible even inside a tunnel, and the image processing is intended to be used for controlling the movement of a moving vehicle inside the tunnel.

FIG. 2 is an overall view of the travel control device according to this embodiment. The moving vehicle 1 includes a video camera 2 and a travel controller 5.
, a sensor 3, and a vehicle controller 4. This mobile vehicle 1 is driven by a normal gasoline engine (not shown) and is steered by a steering mechanism (not shown).

The vehicle controller 4 controls the speed of the vehicle by controlling a throttle actuator, the direction of travel of the vehicle by controlling a steering actuator, and the direction of travel of the vehicle by controlling a brake actuator (not shown). Perform braking. The sensor 3 detects physical information representing the running status of the moving vehicle, and is, for example, a speed sensor, a lateral acceleration sensor, a tire rotation speed sensor, or the like. 6 is a sensor for detecting the imaging angle α of the camera 2. The traveling controller 5 has an image processing device 10 (see FIG. 5) installed therein for recognizing a tangled edge. The travel control for this mobile vehicle is
The left-hand end, the right-hand end, or the center line is detected, and a travel trajectory is generated from the left-hand end, right-hand end, or center line, and travel control is performed. That is, as will be described later, if a tangled edge is detected while traveling outside a tunnel or inside a tunnel, travel control is performed according to a common cruising control algorithm based on the detected tangled edge. In other words, the algorithm for driving control is the same between when the vehicle is traveling outside the tunnel and when it is traveling inside the tunnel, with the only difference being the method of detecting the tangled edge. In addition, with regard to the method of detecting knotted edges, the only difference is the target area of image processing when driving outside a tunnel and while driving inside a tunnel, and the basic tied edge detection algorithm is the same. We are trying to make it smaller.

FIG. 3 shows an image taken by camera 2 outside the tunnel. In the figure, 22 indicates an imaging range. 21 is the center line of the camera, which faces the optical axis direction of the lens.

Further, 20 is an infinity line (horizontal line). In the image,
Unless there is an obstacle in front, the extension line of the left end 23 and the extension line of the right end 25 intersect, and the intersection point P (vanishing point = va
nishingpoint ) will now lie on the horizontal line 20 . The horizontal line 20 may be an actual horizontal line on the earth's surface, or may be a virtual vanishing line drawn horizontally to the vanishing point P.

Figure 4A shows an image obtained inside the tunnel. This tunnel is assumed to have two lanes for one-way travel, and the vehicle is currently traveling in the center of the tunnel. In the figure, point P is a vanishing point, and if the tunnel is a straight road, it is the tunnel exit.

Since the vehicle is running in the center, the vanishing point P appears at the center in the horizontal direction of the image range 22. 31a and 31b are side roads provided near the ends of the tunnel.

30a and 30b are sodium lamps installed on the ceiling inside the tunnel. As mentioned above, because the wavelength range of sodium lamps is biased, side roads are ideal for recognizing tangled ends, but they are useless inside tunnels. In these five practical examples, the sodium lamps 30a and 30b are used as a means of detecting tied ends.

This is because sodium lamp images have a high contrast with other images and are therefore easy to recognize. In this case, as described later, in order to recognize the position of the sodium lamp 30,
The image area above 6 is limited and used as an image processing target area for recognizing the sodium lamp 30. This is to increase processing speed and improve recognition accuracy by limiting the target area.

Figure 1 shows a sodium lamp 30 recognized in this way.
The manner in which the left and right (imaginary) binding ends 34 and (imaginary) center line 35 are detected from the position shown in FIG. These virtual binding ends are obtained from the projection of the image positions of the left and right sodium lamps 30a, 30b obtained above onto the road surface. Further, the position of the virtual center line 35 is determined from the projection of the center point Q between the sodium lamps 30a and 30b onto the road surface.

If the horizontal coordinate values of each of the sodium lamps 30a and 30b in the image are u, then the coordinate value of the center point Q is k. This value accurately points to the center of the tunnel, regardless of whether the vehicle is currently traveling to the right or to the left in the tunnel. Therefore, if this point Q is projected onto the road, it should become the center line 35. Therefore, traveling control may be performed using this virtual center line 35 as a target point.

Incidentally, 32a and 32b are white lines actually drawn on the road to indicate the tangled ends.

The respective positions of the sodium lamps 30a and 30b can be used for detection at the left and right ends. In FIG. 1, the left and right hypothetical ends thus obtained are designated as 34a.

34b. In this case, the position of mounting the sodium lamp 30 on the tunnel wall becomes a problem. sodium lamp 30
is attached close to the center, the virtual connecting end 34a
, 34b are also located at the center, and if running control is performed with this virtual end as the target, there is a possibility that the moving vehicle 1 will continue to run on the center line.

Therefore, the following method is used to selectively use whether the travel control is performed based on the virtual left-bound end 34a or the virtual center line 35, depending on the conditions. That is, the distance 1f between the virtual connecting end 34a and the virtual center line 35
Since Jf t is 1, f2L, for example, compared with a predetermined threshold value S determined from the vehicle width, etc., if t's, travel control is performed with the virtual left-bound end 34a as the target, and t If ≦5, driving control is performed with the virtual center line 35 as the target.

FIG. 5 is a block diagram showing the configuration of an image processing apparatus according to this embodiment. 10 is a frame buffer 10;
Images taken by camera 2 are stored as RGB signals. 12 is a light calculation unit that calculates the amount of light in an image;
GB signal for example YIQ (Y is luminance, IQ is chromaticity signal)
The light intensity is determined from the Y component. If the light level is extremely low and the entire image has a reddish color, the IQ
When there are many acid component signals, it is determined that the tunnel is inside. The vanishing point calculation unit 11 is provided to calculate a vanishing point P and obtain a horizontal line. The vanishing point P is uniquely determined from the camera installation angle α and the camera installation height d. That is, camera 2
If the focal length of is f, then the height position of the vanishing point P within the image is determined as f-tanα. The upper limit switching switcher 13 divides the image vertically along the horizontal line along the virtual horizontal line 36 calculated by the vanishing point calculation unit 11 from the light amount calculation unit 12, and then divides the image into the tunnel from the light amount calculation unit 12. It is determined whether the upper image is to be processed by the road recognition unit 14 or the lower image is to be processed by the road recognition unit 14, according to the determination signal as to whether the person is present or not.

That is, if it is determined that the vehicle is traveling outside a tunnel, the lower image is sent to the road recognition unit 14. If it is determined that the vehicle is running in a tunnel, the image above the horizontal line is copied to the lower area within the buffer and sent to the recognition unit 14. FIG. 4B shows the copy image thus obtained. In the figure,
30a' and 30b' are the copied images of the sodium lamp.

Regarding the image sent from the switcher 13, the road recognition unit 14 follows the determination signal from the light amount calculation unit 12, and if it is determined that the vehicle is traveling outside a tunnel, it recognizes the edge of the road where the vehicle normally travels. If it is determined that the vehicle is traveling in a tunnel, the sodium lamp 30 is detected. If it is determined that the vehicle is traveling in a tunnel, the detected positions of the sodium lamp 30 will be the virtual ends 34a and 34b. In this case, the recognition unit 14 is provided with color information corresponding to the white line on the road for traveling outside the tunnel, and information regarding the color specific to the sodium lamp 30 for traveling inside the tunnel.

In this way, even when driving in a tunnel where the brightness is insufficient and the entire image is biased towards a specific color, the J recognition unit 14 can detect the virtual tangled edges 34a and 34b, and adjust the center line as necessary. 35 is also detected.

In the description of the above embodiment, the horizontal line was recognized based on the angle α at which the camera 2 was attached. This is based on the fact that, as described above, if the vehicle 1 is traveling on level ground, the horizontal line position is uniquely determined by the camera mounting angle α and the camera mounting height d. Note that even if the running path is not flat, the error in the horizontal line detection position based on it is only 133M larger than the size of the target area for image processing for detecting the sodium lamp 30, so it is difficult to detect the stuck edge. It doesn't affect accuracy much.

In the above embodiments, the description has been given of the detection of a tangled edge when traveling in a tunnel, but the present invention can also be applied to the detection of a tangled edge when driving at night. However, in this case, road lights need to be installed side by side at regular intervals on the left and right sides of the road.

(Effects of the Invention) As explained above, the configuration of the image processing device for a moving vehicle according to the present invention includes an Ic image means for taking an image of the traveling path ahead in the direction of travel, and an Ic image means for taking an image of the traveling path ahead in the direction of travel, and an image processing device for detecting that the amount of light in the outside world is below a certain value. A light amount detecting means to detect, a specifying means for specifying an area above the infinity line in the image, and an output from the light amount detecting means, and when the amount of light in the outside world is below a certain value, the specifying means f', fl? It is characterized by the fact that

On driving roads where the amount of light is insufficient (for example, when driving in tunnels or at night), for example, road lights etc. are installed continuously, so the actual binding ends may not be recognized due to the lack of light. In both cases, the road edge can be detected by processing the image data of road lights, etc. taken at the top of the captured image.

[Brief explanation of the drawing]

Fig. 1 is a diagram explaining the principle of an embodiment according to the present invention, Fig. 2 is a diagram explaining the internal configuration of a moving vehicle used in the embodiment, and Fig. 3 is a diagram showing the outside of the tunnel in the embodiment shown in Fig. 2. 4A is a diagram showing an example of an image taken when driving in a tunnel according to the embodiment shown in FIG. 2; FIG. 4B is a diagram illustrating projection of a sodium lamp image by the switcher 13; FIG. 5 is a block diagram showing the configuration of the image processing device according to the embodiment. In the figure, I...Motor vehicle, 2...Camera, 3...Sensor, 4
... Vehicle controller, 5... Travel controller,
6... Camera angle sensor, 10... Image processing device,
11... Vanishing point calculation section, 12... Light amount calculation section, 13
. . . Up/down switching switcher, 14 . . . Road recognition unit, 2
0... Actual horizon, 22... Imaging range, 23...
・Left entangled end of the fruit, 24...Real center line, 25・
・Actual right-handed end, 30a, 30b...Sodium lamp,
31a, 31b...Survey road, 32a, 32b=-actual white line, 34a, 34b...Virtual road edge, 35...
Virtual center line, 36... is a virtual horizontal line. Figure 3

Claims (1)

[Claims] (1) An imaging means for taking an image of the road ahead in the direction of travel; a light amount detection means for detecting that the amount of light in the outside world is below a certain value; and a specification for specifying an area above the line of infinity in the image. and determining means that receives the output of the light amount detection means and determines the end of the running road based on information of the area above the infinity line designated by the designation means when the amount of light in the outside world is below a certain value. An image processing device for a moving vehicle.