JP3341605B2 - Image processing device for vehicles - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle image processing apparatus, and more particularly to a vehicle image processing apparatus capable of extracting a road from a road image.

[0002]

2. Description of the Related Art As a conventional image processing apparatus for a vehicle, for example, an apparatus described in Japanese Patent Application No. 5-354605 is known. The feature is that, as shown in FIG. 7A, a window is set in advance in a trapezoidal area (a) likely to become a road, as shown in FIG. 7A, as a method for detecting a preceding vehicle or an obstacle, and as shown in FIG. Then, the same color region (c) in the window is extracted using the lower center portion (b) of the window as a reference color. Next, as shown in FIG. 7C, a left straight line (d), a right straight line (e), and a straight line (f) serving as an upper boundary are obtained, and the straight line (f) serving as the upper boundary is determined in the y direction. If the position is lower than the upper bottom (g) of the window set in FIG.
That is, it is determined that there is a preceding vehicle or an obstacle.

[0003]

However, in the conventional image processing apparatus for a vehicle, as shown in FIG. 7, as shown in FIG. 7E, there is a problem that the road area (nu) cannot be accurately extracted, and the preceding vehicle and obstacles cannot be detected. Further, for example, as shown in FIG. 7 (d), when a shadow (R) is imaged parallel to the x-axis, there is a problem that a preceding vehicle or an obstacle is erroneously detected. The present invention has been made in view of the above, and an object of the present invention is to provide an image processing method for a vehicle that can contribute to reduction of erroneous detection of a preceding vehicle or an obstacle even when a shade is captured in a road image. It is to provide a device.

[0004]

According to the first aspect of the present invention,
In order to solve the above problems, a photographing means for photographing a road image ahead of a vehicle, a photographing direction detecting means for detecting a photographing direction to be photographed, a time measuring means for measuring a current date and time, and a current position of the own vehicle are detected. Current position detecting means, and a shading shape estimating means for estimating a shading shape appearing on the photographed road image based on the photographing direction, the current date and time, and the position of the vehicle. First object detection means for detecting an obstacle or a preceding vehicle on the road based on the first object detection means;
A road surface area extracting means for extracting, as a road surface region, a region having a similar color to the reference color of the road surface in the photographed road image; and a second detecting device for detecting an obstacle or a preceding vehicle based on the boundary shape of the extracted road surface region. When the boundary shape of the road surface area is similar to the estimated shade shape, the extracted road surface area is not used for detecting an obstacle or a preceding vehicle, and The point is to extract another road surface area.

According to a second aspect of the present invention, in order to solve the above-mentioned problem, there is provided a shade inclination estimating means for estimating a shade value of a shade appearing on the photographed road image, and If the slope value of the straight line representing the boundary is near the estimated shade value of the shade, the road surface area is not used for detecting an obstacle or a preceding vehicle, and another road surface area is extracted. That is the gist.

According to a third aspect of the present invention, in order to solve the above-mentioned problem, an edge for extracting an edge whose inclination value on the photographed road image is in the vicinity of the estimated inclination value of the shade is described. Comprising extraction means, when the straight line representing the boundary of the extracted road surface area matches the extracted edge,
When the slope value of the straight line representing the boundary of the extracted road surface area is near the estimated shade value of the shade, the road area is not used for detecting an obstacle or a preceding vehicle, and another road surface is not used. The point is to extract a region.

According to a fourth aspect of the present invention, there is provided a photographing means for photographing a road image ahead of a vehicle, a photographing direction detecting means for detecting a photographing direction to be photographed, and a current date and time. A clock means, a current position detecting means for detecting a current position of the own vehicle, and a shading shape for estimating a shading shape appearing on the photographed road image based on the photographing direction, the current date and time, and the own vehicle position. Estimating means, first object detecting means for detecting an obstacle or a preceding vehicle on a road based on the estimated shade shape, and detecting a plurality of detection areas set in the photographed road image. A brightness distribution comparing means for comparing the brightness distribution, a third object detecting means for detecting an obstacle or a preceding vehicle based on a result of the brightness distribution comparison, and a slope value on the photographed road image, Of the estimated shade Edge extraction means for extracting an edge near the threshold value, wherein the third object detection means detects an obstacle or a preceding vehicle when the extracted edge exists in the detection area. The gist is to interrupt or keep the previous detection result.

[0008]

[0009]

According to the first aspect of the present invention, a photographing means for photographing a road image ahead of a vehicle, a photographing direction detecting means for detecting a photographing direction to be photographed, and a clocking means for clocking a current date and time. A current position detecting means for detecting the current position of the vehicle, a shading shape estimating means for estimating a shading shape appearing on the photographed road image based on the photographing direction, the current date and time and the vehicle position, Based on the estimated shade shape,
First object detection means for detecting an obstacle or a preceding vehicle on the road; road surface area extraction means for extracting, as a road surface area, an area having a similar color to the reference color of the road surface in the captured road image; A second object detecting means for detecting an obstacle or a preceding vehicle based on the boundary shape of the road surface area, and when the boundary shape of the road surface area is similar to the estimated shade shape, The extracted road surface area is not used for detecting an obstacle or a preceding vehicle, and another road surface area is extracted. This can contribute to a reduction in erroneous detection of obstacles.

According to the second aspect of the present invention, a shade inclination estimating means for estimating a slope value of a shade appearing on a photographed road image is prepared, and the boundary of the extracted road surface area is represented. If the inclination value of the straight line is near the estimated inclination value of the shade, the road surface area is not used for detecting an obstacle or a preceding vehicle, and another road surface area is extracted. In addition, even when a shadow is photographed in a road image, it is possible to contribute to a reduction in erroneous detection of a preceding vehicle or an obstacle.

According to the third aspect of the present invention, there is provided an edge extracting means for extracting an edge whose inclination value on the photographed road image is in the vicinity of the estimated inclination value of the shade. Then, when the straight line representing the boundary of the extracted road surface area coincides with the extracted edge, the slope value of the straight line representing the boundary of the extracted road surface area is close to the estimated shade value of the shade. In some cases, this road surface area is not used for detecting an obstacle or a preceding vehicle, and another road surface area is extracted. Therefore, even when a shadow is captured in a road image, the preceding vehicle is not detected. And erroneous detection of obstacles can be reduced.

According to the present invention, a photographing means for photographing a road image ahead of the vehicle, a photographing direction detecting means for detecting a photographing direction to be photographed, and a time measuring means for clocking a current date and time. A current position detecting means for detecting the current position of the vehicle, a shading shape estimating means for estimating a shading shape appearing on the photographed road image based on the photographing direction, the current date and time and the vehicle position, Based on the estimated shade shape, the first object detecting means for detecting an obstacle or a preceding vehicle on the road is compared with the luminance distribution of a plurality of detection areas set in the photographed road image. A brightness distribution comparing means, a third object detecting means for detecting an obstacle or a preceding vehicle based on a result of the comparison of the brightness distribution, and a slope value of the estimated shade on the photographed road image. Extract edges near the slope value Prepare an edge extracting means for,
The third object detection means interrupts the detection of the obstacle or the preceding vehicle or holds the previous detection result when the extracted edge exists in the detection area. Even if a shadow is taken in the road image,
This can contribute to a reduction in erroneous detection of a preceding vehicle or an obstacle.

[0013]

[0014]

Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) FIG. 1 is a diagram showing a system configuration of a vehicular image processing apparatus according to a first embodiment of the present invention. In FIG. 1, a vehicle image processing apparatus 1 includes a CCD.
Image input unit 11 that captures a road image using a camera or the like
A vehicle position measuring unit 13 that receives a signal from a GPS satellite to measure a vehicle position, a photographing direction detecting unit 14 that detects a photographing direction of the image input unit 11 using a gyro, and the like,
A clock unit 15 that measures the current date and time, and a shade direction calculation that estimates a shade shape by calculating a shade direction on a screen to be shot by the image input unit 11 based on the shooting direction, the current date and time, and the position of the vehicle. Unit 16, image input unit 11, and shade direction calculation unit 16
An image processing unit 12 for detecting an object such as a preceding vehicle or an obstacle based on an output from a vehicle, and a vehicle control unit 17 for determining a control characteristic of the vehicle using a detection result detected by the image processing unit 12. Is done.

The image input unit 11 faces the front end of the vehicle to capture a road image in front of the vehicle.
It is assumed that a CD camera is mounted. In addition, the vehicle position measurement unit 13 does not detect the vehicle position only by the signal of the GPS satellite, but performs a map matching process for adapting the current position to the map information, and a measurement accuracy of the vehicle position using the vehicle speed. The functions mounted on the vehicle position measuring unit 13 and the photographing direction detecting unit 14 may be configured as a single navigation device.

Next, with reference to FIG. 2, a method of calculating the shade direction by the shade direction calculation unit 16 will be described. The shade direction calculating unit 16 first obtains the azimuth and altitude of the sun position shown in FIG. 2A based on the vehicle position and the time using the shade curve shown in FIG. 2B. be able to. This makes it possible to obtain a shade direction on the photographing screen based on the azimuth angle of the sun obtained based on the position and time of the vehicle and the azimuth angle of the photographing direction, and to estimate the shade shape.

[0017]

Next, in step S20, based on the azimuth angle β of the photographing direction and the azimuth angle α of the shade detected by the photographing direction detecting unit 14, the shade of the shade on the plane photographed by the image input unit 11 is determined. When the slope Ws is obtained, Ws = tan (β−α). Next, in step S30, the image processing unit 12
First converts the entire photographing screen from the coordinate system (x, y) in the photographed image to the coordinate system (xr, yr) on the plane photographed according to the following equation derived from FIG.

Yr = 1-11, where l = H / tan θ, l1 = H / tan (dθ1 + θ), dθ1 = arctan (y / f), xr = 13 tan (dθ3) where l3 = H + 11, dθ3 = arctan (X / f) Then, based on the screen after the transformation of the coordinate system, an edge in the same direction as the shade Ws of the shade obtained in step S20 is extracted. Further, an inclination Wci when each edge in the same direction as the extracted shade Ws of the shade is returned to the coordinate system (x, y) in the captured image is obtained. As a simple method of extracting an edge, a well-known Sobel transform may be used.

Next, in step S40, the image processing unit 1
2 sets a window for detecting the road area shown in FIG. In this window, the left and right ends are set slightly wider than the previous recognition result of the road area, and the upper end is set to a fixed y coordinate Wy. If the road width is known, the window may be fixed. Next, step S50
Then, the image processing unit 12 extracts a region of a similar color in the window using a color near the center of the lower end of the window having a high probability of being a road surface as a reference color.

Next, in step S60, the image processing unit 1
2 fits the left, right, and upper boundary points of the road area as shown in FIG. 7B extracted in step S50 to a straight line. As a straight line fitting method, a well-known Hough transform or a least square method may be used. Here, in step S70, the image processing unit 12 determines in step S30
If there is any of the shades Wci of the shade obtained in step S60 that matches the slope of the straight line that is the boundary of the road area obtained in step S60, as shown in FIG. Considering that the image is extracted by being separated by the shadow, the reference color is reset, and the process returns to step S50. At this time, the reference color is the lowermost color other than the area extracted as the road area near the center of the window.

Next, in step S80, the image processing unit 1
2 detects the preceding vehicle. In general, since the vehicle includes many horizontal edge components, the upper boundary of the road area extracted in step S50 is adapted using a horizontal straight line. Then, the obtained straight line y is set as y = c, and if c <Wy, it is determined that there is a preceding vehicle.

Next, the vehicle control unit 17 controls the image processing unit 12
The control characteristic of the vehicle is determined using the detection result detected by the control unit. In step S30, an expression that matches the edge in the shade direction to a straight line is obtained, and in step S70, it is determined whether or not there is a match with a straight line expression that is a boundary of the road area. Even when the shade of the shade in the captured image is horizontal, it can contribute to reducing false detection of a preceding vehicle or an obstacle.

As described above, the image input unit 11 for photographing the road image ahead of the vehicle, the photographing direction detecting unit 14 for detecting the photographing direction to be photographed, and the clock unit 15 for clocking the current date and time.
And a current position measuring unit 13 for detecting the current position of the own vehicle, and based on the photographing direction, the current date and time, and the position of the own vehicle, calculate a shade shape appearing on the photographed road image in a shade direction calculating unit. 16 and an obstacle or a preceding vehicle on the road is determined based on the estimated shade shape by the image processing unit 1.
The detection by 2 can contribute to a reduction in erroneous detection of a preceding vehicle or an obstacle even when a shade is captured in a road image.

Further, the image processing unit 12 includes an image input unit 11
By extracting a region having a similar color to the reference color of the road surface in the road image captured as a road surface region, and detecting an obstacle or a preceding vehicle based on the boundary shape of the extracted road surface region, When the boundary shape is similar to the estimated shade shape, the extracted road surface area is not used for detecting an obstacle or a preceding vehicle, and another road surface area is extracted. Even when a shade is photographed in the image, it can contribute to reduction of erroneous detection of a preceding vehicle or an obstacle.

Further, the image processing unit 12 includes the image input unit 1
By estimating the slope value of the shade appearing on the road image photographed in step 1, the slope value of the straight line representing the boundary of the extracted road surface area is close to the estimated slope value of the shade. Does not use the road surface area for detection of obstacles or preceding vehicles, and extracts another road surface area.Therefore, even when the road image is shaded, Can be reduced.

Further, the image processing unit 12 calculates the inclination value on the road image taken by the image input unit 11 as:
By extracting edges near the estimated shade value of the shade, if the straight line representing the boundary of the extracted road surface area matches the extracted edge, it represents the boundary of the extracted road surface area When the slope value of the straight line is near the estimated shade value of the shade, this road surface area is not used for detecting an obstacle or a preceding vehicle, and another road surface area is extracted. Even when a shade is photographed in a road image, it can contribute to reduction of erroneous detection of a preceding vehicle or an obstacle.

(Second Embodiment) The system configuration of a vehicular image processing apparatus according to a second embodiment of the present invention is the same as the configuration diagram used in the first embodiment shown in FIG. belongs to. Therefore, the description is omitted. Next, the operation of the vehicle image processing apparatus 1 will be described with reference to FIG. 5 and the flowchart shown in FIG. Note that, as shown in steps S210 to S230, the straight line fitting processing of the inclination of the shade and the edge of the shade in the captured image is the same as that of steps S10 to S30 in the first embodiment. Description is omitted. The following processing is performed in the image processing unit 12.

In step S240, it is determined whether or not the straight line representing the edge of the shade passes through area A or area B. The straight line expression representing the edge of the shade is represented by fi (x) = Sa (x-xs) + ys, and the coordinates of the vertices of the area A are represented by A (x
1, y1), B (x1, y2), C (x2, y2), D
Assuming that (x2, y1), when y2> y1, if either y2> fi (x1)> y1 or y2> fi (x2)> y1, the area A is a straight line representing the edge of the shade. Is determined to pass. The same processing is performed for area B. And areas A and B
If it is determined that the straight line representing the edge of the shade passes through any of the above, the determination of the preceding vehicle or obstacle is interrupted, and the process proceeds to step S270.

In step S270, since the time during which the edge of the shade passes through the area A or the area B when the vehicle is running is short, the previous detection result of the preceding vehicle or obstacle is held. To Thereby, it is possible to reduce erroneous detection when a boundary between a shade and a sun appears in the area A or the area B. On the other hand, step S
At 250, the variance σa of the luminance in the window A and the variance σb of the luminance in the window B are calculated.

Next, in step S260, when a preceding vehicle or an obstacle appears in the window B, σb> σa, so that when the difference between the luminance variances σa, σb becomes σb−σa ≧ constant value, It is determined that a car or an obstacle has been detected.

Accordingly, as shown in FIG. 5, an area A near the lower end of the photographed image having a high probability of being a road surface corresponds to an area B at the center of the photographed image having a high probability that a preceding vehicle or an obstacle appears. By comparing the distributions of the respective luminances, when passing through the edge area A or the area B of the shade, the detection is interrupted, or the previous detection result is retained, so that the shade and the sun are maintained. Can be prevented from being erroneously detected due to the boundary of.

As described above, the image processing unit 12 compares the luminance distributions of a plurality of detection areas set in the road image photographed by the image input unit 11, and based on the comparison result of the luminance distributions, Or detect the preceding vehicle,
By extracting an edge whose inclination value on the photographed road image is close to the estimated inclination value of the shade, if the extracted edge exists in the detection area, an obstacle or a preceding obstacle is detected. Because the detection of vehicles is interrupted or the results of the previous detection are retained, even when shade is captured in the road image, it contributes to the reduction of erroneous detection of preceding vehicles and obstacles. Can be.

[Brief description of the drawings]

FIG. 1 is a diagram showing a system configuration of a vehicular image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a diagram showing the azimuth and altitude of the position of the sun (a) and the shade curve (b).

FIG. 3 is a flowchart for explaining an operation of the vehicle image processing device 1 according to the first embodiment of the present invention.

FIG. 4 is a diagram illustrating conversion from a coordinate system in a captured image to a coordinate system on a plane to be captured.

FIG. 5 is a diagram for explaining an operation of the vehicular image processing device according to the second embodiment of the present invention.

FIG. 6 is a flowchart for explaining the operation of the vehicular image processing device 1 according to the second embodiment of the present invention.

FIG. 7 is a diagram for explaining the operation of a conventional vehicle image processing device.

[Explanation of symbols]

 Reference Signs List 11 Image input unit 12 Image processing unit 13 Current position measurement unit 14 Shooting direction detection unit 15 Clock unit 16 Shadow direction calculation unit

Continuation of front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06T 1/00 330 B60R 21/00 624 G06T 7/20 200

Claims (4)

(57) [Claims] 1. A photographing means for photographing a road image ahead of a vehicle, a photographing direction detecting means for detecting a photographing direction to be photographed, a time measuring means for measuring a current date and time, and a current detecting means for detecting a current position of the own vehicle. A position detecting unit; a shading shape estimating unit for estimating a shading shape appearing on the photographed road image based on the photographing direction, the current date and time, and the position of the vehicle; and a shading shape estimating unit based on the estimated shading shape. first and object detecting means, the reference color and similar colors Ryo of the road surface of the captured in road image for detecting an obstacle or a preceding vehicle on the road
Road area extraction means for extracting the area as a road area, and an obstacle or a destination based on the boundary shape of the extracted road area.
Second object detection means for detecting a running vehicle , wherein the boundary shape of the road surface area is similar to the estimated shade shape.
If they are similar, the extracted road surface area is
Extract another road surface area without using it for vehicle detection
An image processing device for a vehicle. 2. A shade appearing on the photographed road image.
A shade gradient estimating means for estimating the gradient value of the, the gradient value of a straight line representing the boundary of the extracted road surface area,
If it is near the estimated shade value of the shade,
Do not use the road surface area for detecting obstacles or preceding vehicles,
The method according to claim 1, wherein another road surface area is extracted.
2. The vehicle image processing device according to 1. 3. A value of a slope on the photographed road image.
Is an edge near the estimated shade value of the shade.
An edge extracting means for extracting a straight line representing a boundary of the extracted road surface area;
The boundary of the extracted road surface area
Is the slope value of the estimated shade value of the shade.
When the vehicle is in the vicinity, the road surface area is
It is not used for both detections, and it is necessary to extract another road surface area.
The vehicle image processing device according to claim 1, wherein: 4. A photographing means for photographing a road image ahead of a vehicle.
When the photographing direction detection means for detecting a photographing direction that is captured, and counting means for counting the current time, the current position detecting means for detecting a current position of the vehicle, the photographing direction, the current date and time and the vehicle position Based on the previous
Day for estimating the shadow shape appearing on the road image taken
Based on the shadow shape estimating means and the estimated shadow shape,
First object detection means for detecting a preceding vehicle, and a plurality of detection areas set in the photographed road image
Brightness distribution comparing means for comparing the brightness distributions of the obstacles or the preceding vehicle based on the comparison result of the brightness distributions
A third object detecting means for detecting the inclination, and a slope value on the photographed road image,
Edge extraction to extract the edge near the shade value of the shade
Output means, and the third object detection means, when the extracted edge exists in the detection area,
Suspend detection of harmful object or preceding vehicle or previous detection
An image processing apparatus for a vehicle, which stores a result.