JP2002148356A - Vehicle outside monitoring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a vehicle outside monitoring device capable of detecting the wet condition of a road surface with a good accuracy. SOLUTION: A wet road determining part 12c determines that the road surface has high possibility of wet road surface when the luminance dispersed value VAR calculated by a luminance dispersed value calculating part 12b is large, and an under road surface data number J calculated by a distance data count part 12a is large. In the wet road determining part 12c, the road surface determined to have high possibility of wet road surface is further determined (inspected) according to the lane reliability D calculated by a road recognizing part 11 and the luminance ratio (Bl/Br) of a white line to the road surface, and when the lane reliability D is low, or when the luminance ratio (Bl/Br) is small, the road surface is determined to be a wet road surface.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an external monitoring system for detecting a road surface condition of a traveling road, particularly a wet road surface, based on a captured image.

[0002]

2. Description of the Related Art In recent years, attention has been paid to a vehicle exterior monitoring device using a vehicle-mounted camera having a solid-state imaging device such as a CCD.
This device recognizes a driving environment based on an image captured by a vehicle-mounted camera, calls attention to a driver, and performs vehicle control such as downshifting as necessary.

[0003] Incidentally, in this type of vehicle exterior monitoring device, there is one that detects a road surface condition based on a captured image and reflects the detected road surface condition in vehicle control or the like.
For example, Japanese Patent Application No. 11-216713 filed by the present applicant discloses that on a road surface wet by rain or the like, a three-dimensional object is reflected on the road surface to detect the wet state of the road surface, and the wet state of the road surface is detected. In such a case, a technique for temporarily interrupting vehicle control or changing control parameters is disclosed. According to this technology, the distance data is calculated based on the parallax of the object in the pair of image data, and the three-dimensional shape of the road is recognized based on the image data and the distance data. Also recognizes the wet state of the road surface based on the number of distance data (the number of wet data) existing below.

[0004]

By the way, on a road surface, exceptionally, even when the road surface is not in a wet state,
The distance data may be detected below the road surface position. As an exceptional case, for example, as shown in FIG. 8, there is a case where headlight light of an oncoming vehicle is reflected on a road surface at the time of running at night or the like. In such a case, the road surface may be detected as being wet even though the road surface is dry.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a vehicle exterior monitoring device that can accurately detect a wet state of a road surface.

[0006]

According to a first aspect of the present invention, there is provided an out-of-vehicle monitoring device for detecting a road surface condition from a captured image based on the image. A road recognition unit that recognizes a road shape including lane detection; and a road surface state recognition unit that determines whether the road surface of the road recognized by the road recognition unit is wet. The means determines the wet state of the road surface based on the lane state detected by the road recognition means.

According to a second aspect of the present invention, in the vehicle outside monitoring apparatus according to the first aspect, the road surface state recognizing means is configured to detect a reliability of the lane detected by the road recognizing means by a value greater than a set threshold value. When low, it is determined that the road surface is in a wet state.

According to a third aspect of the present invention, there is provided a vehicle exterior monitoring device according to the first or second aspect.
The road surface state recognizing means determines that the road surface is wet when the luminance ratio between the lane and the road surface on the road is smaller than a set threshold.

According to a fourth aspect of the present invention, there is provided a vehicle exterior monitoring apparatus according to any one of the first to third aspects, based on a pair of captured images, the distance between the object on the images. Stereo image processing means for calculating data, wherein the road recognition means performs three-dimensional recognition of a road shape including lane detection based on the distance data and the luminance information of the image; The means calculates the number of the distance data existing below the road surface position in the setting area on the road, and when the number of the distance data existing below the road surface position is smaller than the set value. It is characterized in that it is determined that the road surface is not in a wet state.

According to a fifth aspect of the present invention, in the vehicle exterior monitoring apparatus according to any one of the first to fourth aspects, the road surface state recognizing means includes a road surface within an area set on the road. Means for calculating the luminance variance value of the road surface, and when the luminance variance value of the road surface is smaller than a set value, it is determined that the road surface is not in a wet state.

[0011]

Embodiments of the present invention will be described below with reference to the drawings. The drawings relate to an embodiment of the present invention, and FIG. 1 is a functional block diagram of a stereo exterior monitoring device.
FIG. 2 is an explanatory diagram showing a lane detection area on an image, and FIG.
4 is an explanatory diagram of a lane model, FIG. 4 is an explanatory diagram showing a distance data monitoring area, FIG. 5 is a flowchart showing a wet road determination routine, FIG. 6 is an explanatory diagram showing an example of an image when driving on a dry road surface, and FIG. FIG. 8 is an explanatory diagram illustrating an example of an image when traveling on a road surface, and FIG. 8 is an explanatory diagram illustrating an example of an image when traveling on a dry road surface at night.

In FIG. 1, a pair of cameras 1 and 2 having a built-in image sensor such as a CCD are mounted at predetermined intervals in a vehicle width direction of a vehicle such as an automobile, and capture images of a scene in front of the vehicle. The main camera 1 captures a reference image (right image) necessary for performing stereo processing, and the sub camera 2 captures a comparison image (left image) in this processing. In a state where the cameras 1 and 2 are synchronized with each other,
Each of the analog images output from 2 is converted into a digital image of a predetermined luminance gradation (for example, 256 gray scales) by A / D converters 3 and 4. The digitized image is subjected to luminance correction, geometric conversion of the image, and the like in the image correction unit 5. Usually, since the mounting positions of the pair of cameras 1 and 2 have an error to some extent, a shift caused by the error exists in the left and right images. In order to correct this displacement, geometric transformation such as rotation or translation of the image is performed using affine transformation or the like. The reference image and the comparative image corrected in this way are stored in the original image memory 8.

On the other hand, a stereo image processing section 6 as a stereo image processing means uses the reference image and the comparison image corrected by the image correction section 5 to convert the three-dimensional position of the same object in the image (from the own vehicle to the object). Is calculated). This distance can be calculated based on the principle of triangulation from the relative displacement of the position of the same object in the left and right images. The distance information of the image thus calculated is stored in the distance data memory 7.

The microcomputer 9 has a three-dimensional object recognizing unit 10, a road recognizing unit 11 as a road recognizing unit, and a road surface state recognizing unit 12 as a road surface recognizing unit. 8 and distance data memory 7
Based on each piece of information stored in the vehicle, recognition of a three-dimensional object such as a traveling vehicle ahead of the vehicle, recognition of a road ahead of the vehicle including detection of a lane (white line, etc.), recognition of a road surface state on the recognized road, and the like are performed. Here, as the road surface state recognition performed by the road surface state recognition unit 12, it is determined whether or not the road surface is in a wet state (asphalt wet road surface, hereinafter also simply referred to as a wet road surface). Then, these recognition results are input to the processing unit 13, and based on each input information, the processing unit 13 alerts the driver with an alarm device 19 such as a monitor speaker when an alarm is required, Alternatively, the control units 14 to 18 are controlled as needed. For example, AT
(Automatic transmission) Instructs the control unit 14 to execute downshifting. Further, it may instruct the engine control unit 18 to reduce the engine output. In addition, the antilock brake system (ABS) controller 1
5. Appropriate vehicle control can be instructed to the traction control system (TCS) control unit 16 or the vehicle behavior control unit 17 that controls the torque distribution and the number of revolutions of each wheel.

The road recognition unit 11 recognizes a road shape by a method described in Japanese Patent Application No. 11-269578 filed by the present applicant, for example. That is, the road recognizing unit 11 sets a reference image having a 512 × 200 pixel area within a search range that is set in advance or variably set according to the pitching state of the host vehicle (search start line js to Lane detection is performed for each horizontal line of the search end line je (see FIG. 2), and based on the detected lane position and distance information from the distance data memory 7, lane position recognition in real space (that is, road recognition). I do.

More specifically, the road recognizing unit 11 first calculates the road surface luminance Aroad for each search line j. The road surface luminance Aroad is obtained in the left and right regions excluding the central portion of the image. In the search start line js, the search start line js and four pre-horizontal lines jpre before the search start line js are obtained. Is performed on the basis of the luminance information that has already been detected in the subsequent search lines.

The road recognition unit 11 performs lane detection for each target line j based on a luminance determination threshold value calculated from the road surface luminance Aroad, luminance information of an image, distance information, and the like. That is, the road recognition unit 11 determines the lane start point and the end point based on the edge strength (luminance differential value), and determines the lane start point and the end point based on the comparison between the luminance determination threshold and the luminance. Extract candidates. Then, based on the distance information, by verifying whether or not the lane candidate is on the road surface, only the lane is extracted from the lane candidate points. Here, the lane start point refers to a lane point corresponding to the inner boundary between the lane and the asphalt, and the lane end point refers to a lane point corresponding to the inner boundary between the lane and the asphalt.

Further, the road recognizing unit 11 calculates the lane point P (i, j) obtained by the lane detection and the parallax d at that point.
(I, j, d) are obtained for all lane points P, and the positions (X, Y, Z) of the left and right lanes in the real space are uniquely calculated based on the camera specifications and the like. Then, in the road recognizing unit 11, a function representing the road shape (road model)
By setting each of the parameters so as to match the road shape, the predicted travel line L (see FIG. 3) is calculated. That is, in the example of FIG. 3, the recognition range is set to a predetermined distance (Z1 to Z1).
Z7) is divided into seven sections, and the lane point P in each section is linearly approximated by the least squares method. (Lane model) Horizontal shape model Left lane X = aL · Z + bL Right lane X = aR · Z + bR Road height model Left lane Y = cL · Z + dL Right lane Y = cR · Z + dR And the left and right sides calculated in this way The predicted travel line L can be calculated from the line.

Further, the road recognition unit 11 considers the presence / absence (number) of the detected lane points P and the reliability of the left and right lanes in consideration of the continuity with the lane points P detected in the previous frame. Calculate D. In this case, each lane reliability D is
It becomes high when the lane point P is continuously detected.

Further, the road recognizing unit 11 calculates the average lane luminance Bl and the average road surface Br on the recognized road, and obtains the luminance ratio (Bl / Br). Here, it is desirable that the road surface area when calculating the road surface average luminance is set to be a horizontally long area in order to reduce the influence of the luminance change in the front-rear direction. In addition, in consideration of running at night, it is desirable that the road surface area is an area that is within the irradiation range of the own vehicle light and is less affected by disturbance due to other illumination light or the like. Therefore, in the present embodiment, the lane average luminance Bl and the road average luminance Br are calculated for the road of the portion detected in the second area (see FIG. 3) of the road model, and the luminance ratio (Bl /
Br). In this case, the calculation of the lane average brightness Bl is performed using only the highly reliable one in which the lane point P is continuously detected from the previous frame.
The calculation of the road surface average brightness Br is performed on the inner road surface of the detected left and right lanes in order to reduce the influence of roadside objects.

The road surface condition recognizing unit 12 detects distance data (under-road data) existing below the road surface of the recognized road and counts the number J of data under the road surface, and a distance data counting unit 12a. A brightness variance value calculation unit 12b that calculates a brightness variance value VAR on the road surface of the recognized road;
A wet road determination unit 12c that determines a wet state of the road surface based on the number J of data under the road surface, the luminance dispersion value VAR, the lane reliability D, the luminance ratio (Bl / Br), and the like. .

The distance data count section 12a is provided with a data number J on the road surface and a data number J under the road surface by the method described in Japanese Patent Application No. 11-216713 by the present applicant.
Is to count.

That is, in the distance data counting unit 12a, first, the predicted travel line L is input from the road recognition unit 11, and as a reference of the predicted travel line L, for example, 40 m in the forward direction of the vehicle in the corresponding real space. 0 ≦ Z ≦ 4
0), and a distance data monitoring region R of 2 m (-2 ≦ X ≦ 2) is set in each of the left and right width directions (see FIG. 4).

The distance data counting section 12a specifies effective distance data from the distance data existing in the distance data monitoring region R. Here, the effective distance data is calculated, for example, in units of 4 × 4 pixel blocks, and refers to distance data relating to a pixel block having a predetermined number or more of luminance edges in the horizontal direction (horizontal direction) of an image.

Further, the distance data counting unit 12a classifies each of the specified effective distance data into three items, three-dimensional object data, on-road data, and under-road data, and counts the number of each data. Here, the three-dimensional object data is distance data calculated due to a three-dimensional object such as a preceding vehicle,
Distance data of Y> 0.3 (unit: m) is classified as three-dimensional object data. The road surface data is distance data calculated based on the road surface (lane, road surface rut, gravel, etc.) of the traveling road, and the distance data of −0.4 ≦ Y ≦ 0.3 is calculated on the road surface. Classified as surface data. The under-road data is distance data calculated due to reflection of a three-dimensional object on a road surface wet by rain, and Y <
The distance data of 0.4 is classified as under-road data.

The luminance variance value calculating section 12b calculates the luminance variance value VAR by a method described in detail in, for example, Japanese Patent Application No. 11-216915 filed by the present applicant. In this case, the brightness variance value calculation unit 12b does not calculate the variance value for the brightness over the entire image, but calculates the brightness variance value VAR limited to the recognized area on the road.

That is, in the luminance variance value calculation unit 12b,
A rectangular image area corresponding to the road recognized by the road recognition unit 11 is set on the original image, and the image area is divided into predetermined pixels in the horizontal direction and monitored in a rectangular shape (extending in the vertical direction of the image). An area Ni (1 ≦ i ≦ n) is set.

Then, the luminance variance calculating unit 12b calculates the luminance sum A for each monitoring area N. The luminance sum Ai in a certain monitoring region Ni can be calculated as a value obtained by extracting a plurality of pixels uniformly distributed in the region as a sample and adding their luminance values (or their average value). .

Then, the brightness variance value calculation unit 12b calculates the variance value VAR of the brightness distribution characteristics of the image area based on the following equation. Here, n is the number of monitoring areas, and Aave is the average value of the luminance sum.

The wet road determining unit 12c calculates the lane reliability D calculated by the road recognizing unit 11 and the luminance ratio (Bl) between the lane and the road surface.
/ Br) is input, and the number J of data under the road surface counted by the distance data counting unit 12a and the luminance variance value VAR calculated by the luminance variance value calculation unit 12b are input. Based on these pieces of information, Judge the wet road surface. The wet road determination unit 12c turns on the wet road surface determination flag F when it is determined that the vehicle is on a wet road surface.
Otherwise, the WET road surface determination flag F is turned off.
I do. Here, in addition to the above information, the own vehicle speed from a vehicle speed sensor (not shown) and information on the preceding vehicle from the three-dimensional object recognition unit 10 are input to the wet road determination unit 12c.
Prior to the wet road surface determination, it is determined whether or not the wet road surface determination is currently possible.

In the following, the wet road surface determination process will be described.
This will be described in detail with reference to the flowchart of FIG. This routine is executed every set time, and is executed in step S1.
01, first, the own vehicle speed is set to a preset speed V
It is determined whether it is 1 or more. Here, the set speed V1 is set to a predetermined low speed that the host vehicle will take when traveling on a narrow road or a parking lot. That is, in narrow roads, parking lots, and the like, it is often difficult to obtain sufficient road information based on an image because of a wall in front or no lane, etc. It is difficult to determine a wet road surface. In addition, when driving at such a low speed, even on a wet road surface,
Generally, there is no need to change the vehicle control or to issue a warning.

Therefore, when the own vehicle speed is lower than the set speed V1 in step S101, it is determined that it is difficult to appropriately determine a wet road surface, and step S114 is performed.
Then, after the WET determination flag F is turned off, the routine exits.

On the other hand, if it is determined in step S101 that the own vehicle speed is equal to or higher than the set speed V1,
Proceed to step S102.

In step S102, the three-dimensional object recognition unit 10
Is determined on the basis of the information from the vehicle to determine whether the inter-vehicle distance between the host vehicle and the immediately preceding vehicle is equal to or greater than a preset distance l. Here, the set distance 1 is a minimum inter-vehicle distance so that a road surface area necessary for wet road surface determination can be imaged without being covered by a preceding vehicle, and is a distance set in advance by an experiment or the like. That is, in step S102, it is determined whether the preceding inter-vehicle distance is equal to or greater than the set distance 1 to determine whether or not appropriate wet road surface determination is possible.

If the preceding inter-vehicle distance is smaller than the set distance 1 in step S102, it is determined that it is difficult to image a necessary road surface area and it is difficult to determine a proper wet road surface, and the process proceeds to step S114. After turning off the WET determination flag F, the routine exits.

On the other hand, if it is determined in step S102 that the preceding inter-vehicle distance is equal to or greater than the set distance l, the process proceeds to step S103.

In step S103, it is determined whether or not the road surface luminance variance VAR calculated by the luminance variance value calculator 12b is equal to or greater than a predetermined threshold VAR1. Here, the threshold value VAR1 is a threshold value for distinguishing between an asphalt dry road surface (hereinafter, simply referred to as a dry road surface) and another road surface condition. That is, on a wet road surface, the luminance on the road surface often varies due to reflection of a three-dimensional object or the like, and in such a case, the luminance variance VAR of the captured road generally shows a high value. Similarly, on a uneven snowy road surface where snow partially remains on the road surface or a road surface made of gravel or the like, the luminance variance value VAR is similarly calculated.
Indicates a high value. On the other hand, a change in luminance is generally small on an asphalt dry road surface or the like, and the luminance variance VAR of a captured image shows a low value. Therefore, by appropriately setting the threshold value VAR1 based on the luminance variance value and the like in each road surface state obtained by experiments and the like, and comparing the threshold VAR1 with the calculated luminance variance value VAR, the dry road surface and other road surface conditions are compared. Can be distinguished.

In step S103, the luminance variance value V
If AR is smaller than the threshold value VAR1, it is determined that the road surface is likely to be a dry road surface, and the process proceeds to step S114.

On the other hand, if the luminance variance value VAR is equal to or larger than the threshold value VAR1 in step S103, it is determined that there is a high possibility that the road surface is other than a dry road surface, and the flow advances to step S104.

In step S104, it is determined whether or not the number J of data under the road counted by the distance data counting unit 12a is equal to or greater than a predetermined threshold value J1. Here, the threshold value J1 is a threshold value for determining between a wet road surface and another road surface condition. That is, although there is not much difference in the luminance variance VAR between the wet road surface and the uneven snow road surface or the road surface made of gravel or the like, a difference occurs in the number of distance data under the road surface (the number J of data under the road surface). . That is, on a wet road surface, the variance of luminance is mainly due to reflection of a three-dimensional object on the road surface, and therefore, the number of distance data under the road surface (the number J of data under the road surface) increases. On the other hand, on an uneven snowy road surface or a road surface made of gravel or the like, since the dispersion of luminance is caused by a pattern on the road surface, the number J of data under the road surface becomes small. Therefore, by appropriately setting the threshold value J1 based on the number of data under the road surface in each road surface state obtained by experiments and the like, and comparing the threshold value J1 with the calculated number J of data under the road surface, the wet road surface and the other road surface are compared. Can be distinguished.

In step S104, when the number J of data under the road surface is smaller than the threshold value J1, it is determined that there is a high possibility that the road surface is formed of an uneven snow condition or gravel, and the flow advances to step S114.

On the other hand, if the number J of data under the road surface is equal to or larger than the threshold value J1 in step S104, it is determined that there is a high possibility that the vehicle is on a wet road surface, and the flow advances to step S105.

Subsequent steps S105 and S10
No. 6 performs further determination (verification) on the road surface determined to be highly likely to be a wet road surface in steps S103 and S104 described above. That is,
As described above, the wet road surface (see FIG. 7) has the characteristics that the luminance variance value VAR is large and the number J of data under the road surface is large.
By the process of S104, it can be distinguished from a dry road surface (see FIG. 6) and a road surface made of uneven snow or gravel.
However, for example, as shown in FIG. 8, even on a dry road surface, when illuminated by the headlights of an oncoming vehicle at night or the like, the luminance variance value VAR of the road surface and the number J of data under the road surface are large. May be indicated. Therefore, in addition to the above-described determination based on the luminance variance value VAR and the number J of data under the road surface, the wet road determination unit 12c performs steps S105 and S10.
At 6, a further wet road surface determination based on the lane on the road surface is made.

In step S105, it is determined whether the lane reliability D of at least one of the left and right lanes is equal to or greater than a predetermined threshold value D1. Here, the threshold value D1 is set to the lane reliability that would normally be obtained on a dry road surface. That is, in general, on a dry road surface, the luminance difference between the lane and the road surface and the edge strength (luminance change amount) are larger than on the wet road surface, so that it is easy to continuously detect the lane, and the lane reliability is improved. D increases. Therefore, it is possible to appropriately set the threshold value D1 based on the lane reliability of each road surface state obtained by experiments and the like, and to compare the calculated lane reliability D with the dry road surface and the wet road surface. it can.

In step S105, when both the left and right lane reliability D are smaller than the threshold value D1, the road surface is determined to be a wet road surface, and the flow advances to step S109. As shown in FIG. 8, even on a dry road surface,
The detection of the lane may be temporarily difficult due to the headlight light of the oncoming vehicle, but since the influence of the headlight light or the like is partial, the lane reliability D due to such light or the like may be caused. It is difficult to imagine that the decrease in the vehicle speed occurs simultaneously in the left and right lanes. Therefore, in step S105, the process proceeds to step S109 only when both the left and right lane reliability D are smaller than the threshold value D1.

On the other hand, if it is determined in step S105 that the lane reliability D of at least one of the left and right lanes is equal to or greater than the threshold value D1, the process proceeds to step S106.

In step S106, the lane average luminance Bl
It is determined whether or not the luminance ratio (Bl / Br) of the road surface average luminance Br is equal to or greater than a preset threshold (Bl / Br) 1. Here, the threshold value (Bl / Br) 1 is set to the lowest luminance ratio between the lane and the road surface that would normally be obtained on a dry road surface. That is, the difference in brightness between the lane and the road surface is generally smaller on a wet road surface than on a dry road surface. Therefore, the threshold value (Bl / Br) 1 is appropriately set based on the luminance ratio of each road surface condition obtained by experiments and the like, and is compared with the calculated luminance ratio (Bl / Br), whereby the wet road surface and the wet road surface are wetted. It can be distinguished from the road surface.

In step S106, the luminance ratio (Bl
If (/ Br) is smaller than the threshold value (Bl / Br) 1, the road surface is determined to be in a wet state, and the process proceeds to step S109.

On the other hand, if the luminance ratio (Bl / Br) is equal to or more than the threshold value (Bl / Br) 1 in step S106, it is determined that the road surface is in a dry state, and step S10 is performed.
Go to 7.

In step S107, WET indicating the number of times that the vehicle has been judged to be a wet road surface by the above-described processing.
It is checked whether or not the road surface counter Cwet is larger than a preset counter minimum value Cmin. ← Cwet-1), and then step S11
Proceed to 1.

On the other hand, in step S107, WET
When the road surface counter Cwet is the counter minimum value Cmin, the process directly proceeds to step S111.

Step S105 or S106
In step S109, it is determined whether or not the wet counter Cwet is smaller than a preset maximum counter value Cmax. If the wet counter Cwet is smaller than the maximum counter value Cmax, the process proceeds to step S110. To WET road counter Cwe
After incrementing t (Cwet ← Cwet + 1), the process proceeds to step S111.

On the other hand, in step S109, WET
If the road counter Cwet is equal to the counter maximum value Cmax, the process proceeds to step S111.

In step S111, the WET road surface counter Cwet is set to a predetermined threshold value Cwet1 (Cmin <Cwet1 <
Cmax) or more is checked, and if the WET road surface counter Cwet is equal to or more than the threshold value Cwet1, step S11 is performed.
Proceed to 2. Then, in step S112, a wet road surface determination flag F indicating that the road surface is a wet road surface
Is turned on (F ← 1), and the routine exits.

On the other hand, in step S111, WET
When the road surface counter Wet is smaller than the threshold value Wet1, the process proceeds to step S113.

In step S113, the wet road surface counter Cwet is set to a predetermined threshold value Cwet2 (Cmin <Cwet2 <
It is checked whether or not Cwet1 <Cmax) or less. If the WET road surface counter Cwet is equal to or less than the threshold value Wet2, the process proceeds to step S114, and the WET road surface determination flag F is turned off.
After (F ← 0), the routine exits.

On the other hand, in step S113, WET
If the road counter Cwet is larger than the threshold value Cwet2, the routine exits from the routine.

In such an embodiment, the road shape including lane detection is recognized based on the captured image, and it is determined whether the road surface is a wet road surface based on the detected lane condition. Therefore, the wet state of the road surface can be accurately detected.

That is, it is generally difficult to continuously detect a lane on a wet road surface due to a decrease in luminance difference between the lane and the road surface, a decrease in edge strength, and the like, as compared with a dry road surface. Focusing on the fact that the detected lane reliability decreases, it is determined that the road surface is a wet road surface when the lane reliability is low, so that it is possible to accurately detect that the road surface is a wet road surface. Furthermore, on a wet road surface, it is generally noted that the luminance difference between the lane and the road surface is smaller than that on a dry road surface, and when the luminance ratio is small, it is determined that the road surface is a wet road surface. Can be accurately detected.

In other words, by judging that the road surface is a wet road surface based on the condition of the lane, it is erroneously determined that the road surface is wet when headlight light of an oncoming vehicle is reflected on a dry road surface. The determination can be prevented.

Further, by calculating the luminance variance value of the road surface and determining that the road surface is likely to be a dry road surface when the luminance variance value of the road surface is small, the determination accuracy of the wet road surface is further improved. Can be.

Further, the number of distance data under the road surface (the number of data under the road surface) is calculated, and when the number of data under the road surface is small, it is determined that there is a high possibility that the road surface is other than the wet road surface. The determination accuracy can be further improved.

In the above embodiment, each threshold value may be variably set according to the driving environment. In this case, for example, the accuracy of the wet road surface determination can be further improved by variably setting each threshold to an appropriate value according to the shutter speed of the cameras 1 and 2.

[0064]

As described above, according to the present invention, the road shape including the lane detection is recognized, and the wet state of the road is determined based on the detected state of the lane. It can be detected well.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a stereo exterior monitoring device.

FIG. 2 is an explanatory diagram showing a lane detection area on an image.

FIG. 3 is an explanatory diagram of a lane model.

FIG. 4 is an explanatory diagram showing a distance data monitoring area.

FIG. 5 is a flowchart showing a wet road determination routine;

FIG. 6 is an explanatory diagram showing an example of an image when traveling on a dry road surface.

FIG. 7 is an explanatory diagram showing an example of an image when traveling on a wet road surface.

FIG. 8 is an explanatory diagram showing an example of an image when driving on a dry road surface at night.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Main camera 2 Sub camera 5 Image correction part 6 Stereo image processing part (stereo image processing means) 7 Distance data memory 8 Original image memory 9 Microcomputer 10 Three-dimensional object recognition part 11 Road recognition part (road recognition means) 12 Road surface state recognition Unit (road surface state recognition means) 12a distance data counting unit 12b luminance variance value calculation unit 12c wet road determination unit

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) B60R 21/00 626 B60R 21/00 626A G01B 11/00 G01B 11/00 H 11/24 11/24 K

Claims (5)

[Claims] An exterior monitoring device for detecting a road surface condition of a road from a captured image, comprising: a road recognizing means for recognizing a road shape including lane detection based on the image; Road condition recognizing means for determining whether or not the road surface of the road is in a wet condition, wherein the road condition recognizing device determines the wet condition of the road surface based on the lane condition detected by the road recognizing device. A vehicle exterior monitoring device characterized in that: 2. The road surface condition recognizing device determines that the road surface is wet when the reliability of the lane detected by the road recognizing device is lower than a set threshold value. The vehicle exterior monitoring device according to claim 1. 3. The road surface condition recognizing means determines that the road surface is in a wet state when the luminance ratio between the lane and the road surface on the road is smaller than a set threshold value. The vehicle exterior monitoring device according to claim 2. 4. A stereo image processing means for calculating distance data of an object on the image based on a pair of captured images, wherein the road recognizing means includes: the distance data; the luminance information of the image; Based on the three-dimensional recognition of the road shape including lane detection, the road surface state recognizing means calculates the number of the distance data existing below the road surface position in the set area on the road. The road surface is determined not to be in a wet state when the number of distance data present below the road surface position is smaller than a set value. Outside monitoring device. 5. The road surface condition recognizing means includes means for calculating a luminance variance value of a road surface in an area set on the road, and when the luminance variance value of the road surface is smaller than a set value, the road surface becomes wet. 2. The method according to claim 1, wherein the state is determined not to be a state.
The vehicle exterior monitoring device according to claim 4.