JP2002236912A - Road partition line recognizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a road partition line recognizing device capable of estimating a road parameter with high accuracy. SOLUTION: A partition line candidate line becoming a candidate for a partition line is detected from an image photographed by a CCD camera 3, image coordinates of a representative point set on the partition line candidate line are calculated, and inclination of the partition line candidate line in each representative point is obtained. According to a first equation containing the respective parameters C, A, B, D and H, the image coordinates of the respective representative points and the inclination of the partition line candidate line, the product Z of the parameter B by H (=B*H) and the parameters C, D are estimated. Subsequently, according to a second equation containing the past value of the product Z, the parameter B is obtained, and two or more representative points are selected for one of the left partition line candidate line and the right partition line candidate line. According to the image coordinates of the respective representative points, the parameters B, C, D and a third equation containing the parameters C, A, B, D and H, the parameters A and H are estimated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a partition line recognition device for accurately recognizing a partition line such as a white line drawn on a road.

[0002]

2. Description of the Related Art As a conventional white line (partition line) recognition device,
Japanese Patent Application Laid-Open No. Hei 6-20189, "Road Shape Measuring Apparatus" (hereinafter referred to as Conventional Example 1) and Japanese Patent Application Laid-Open No. 8-261756, "Running Lane Recognition Apparatus" (hereinafter referred to as Conventional Example 2) are known. Have been.

In each of the above conventional examples,
There is a possibility that a false white line such as a rut, a shadow, a water film, and a double white line appearing along the road white line is erroneously recognized, and the estimation of the road model fails. Hereinafter, a processing procedure by the white line recognition device according to the conventional example 1 will be described with reference to a flowchart shown in FIG.

<Step s101: Capture of image in front of vehicle> Images captured by a camera fixed to the vehicle and prepared for capturing an image in front of the vehicle are captured at predetermined intervals.

<Step s102; Window setting>
FIG. 9 shows an example of the window. In general, a plurality of windows (in the figure, a total of ten windows, five on the left and five on the right) are prepared, and a white line is detected in this window. The position of this window is estimated based on the road model obtained by the image processing result at the time of the previous video capturing.

The road model is represented by the following equation (5).

[0007]

(Equation 4) The above equation (5) is detailed in Conventional Example 1, and details such as derivation of the equation are omitted. Equation (5) is a formulation that considers the lane width as a constant, and there is another formulation that considers the camera height as a constant. In both cases, the number of unknown parameters is five, which is equivalent. Therefore, the following description will be made according to equation (5).

In equation (5), parameters ｛A, B,
C, D, and H} correspond to the lateral position of the own vehicle relative to the lane, the road curvature, the yaw angle, the pitch angle, and the camera height of the own vehicle relative to the lane, respectively. E0 is the lane width (the interval between the white line on the left side and the white line on the right side), and fv and fh are camera perspective conversion constants, which represent the vertical direction and the horizontal direction, respectively. {X, y} is a coordinate image of an arbitrary point on the white line candidate line. When E0 is the distance between the insides of the lanes, {x, y} is a coordinate image of an arbitrary point inside the white line candidate point.

Hereinafter, x and y mean coordinates on an image. i represents 0 on the left white line and 1 on the right white line. N
The y coordinate of the upper side of the second window (smaller y coordinate) is yn. The previous parameter estimation results are expressed as ｛A-1, B-
1, C-1, D-1, H-1}, the center coordinate xnc of the Nth window this time is expressed by the following equation (6).

[0010]

(Equation 5) A method has been proposed in which the width of the window in the x direction has a fixed value, or a method of rationally setting the variance of parameters as shown in Conventional Example 2.

<Step s103; white line candidate point selection> so
Use edge detection (brightness change) such as a bel filter.
As shown in FIG. 10, the left (the smaller the x value of the image coordinates)
If the luminance of the pixel is larger than the luminance of the right pixel, the output of the filter is assumed to be positive. Since the inside of the white line is a white line candidate point, the output of the filter is positive at the white line candidate point for the left white line and negative at the right white line. In the case of the left white line, the images of all the windows are scanned, and those whose output of the filter exceeds a predetermined positive value are regarded as white line candidate points.

<Step s104: White Line Candidate Line Selection> Processing for obtaining a white line candidate line from a set of white line candidate points includes Hough transform, least square method, and the like. For example, in the Hough transform straight line approximation, as shown in FIG. 11, a straight line passing through the window and passing through the most white line candidate points is selected.

<Step s105: Calculation of representative point of white line candidate line> One point passing through the white line candidate line is selected, and its image coordinates are obtained. For example, the intersection of the white line candidate line and the upper side of the window is set as the representative point. This is represented by {xn, y in FIG.
It is equivalent to n｝.

<Step s106: Road model parameter estimation> Set of representative points {x0, y0} to {xn0, yn0}
The road model {A, B, C, D, H} is estimated based on (n0: number of representative points) and the above equation (5). As representatives of this estimation method, there are a least square method and a Kalman filter.

[0015]

Next, problems will be described. In step s106, the road parameters cannot be estimated unless at least one of the left and right points is given as the representative point. Hereinafter, this will be described with an example.

When only the left white line is given, the above equation (5) is rewritten into the following equation (7).

[0017]

(Equation 6) As is clear from the equation (7), even if the parameter H is multiplied by k, the same equation is established if the parameter A is multiplied by k and the parameter B is multiplied by (1 / k). That is, when only the left white line is given, the parameters A, B, and H cannot be determined. The same applies to the case where only the right white line is given.

When using a stereo camera,
Since there are two non-equivalent equations corresponding to equation (7), even if only one white line is given,
Although parameters can be estimated, when a monocular camera is used, only equation (7) is given, and road parameters cannot be estimated.

As a method for solving this problem, there is a method in which the parameter H is assumed to be a fixed value Hs. Although it is possible to estimate the parameters A, B, C, and D by using this method, as described above, when the true value of the parameter H is k times the fixed value Hs, The estimated value of A is k times, and the estimated value of parameter B is (1 /
k) times, and an error occurs.

The present invention has been made to solve such a conventional problem, and its object is to provide:
It is an object of the present invention to provide a road partition line recognition device capable of accurately estimating road parameters even when only one road white line exists.

[0021]

According to a first aspect of the present invention, there is provided a monocular photographing means for photographing a road image in front of a vehicle, wherein the photographing is performed by the photographing means. Based on the image, a parameter C corresponding to the yaw angle, a parameter A corresponding to the lateral displacement, a parameter B corresponding to the curvature, a parameter D corresponding to the pitch angle, and a parameter H corresponding to the camera height are estimated and estimated in discrete time. In the road partition line recognition device that recognizes the partition line included in the road image based on the respective parameters, a partition line candidate line serving as a partition line candidate is detected from the image captured by the image capturing unit. Means for calculating image coordinates of a representative point set on the partition line candidate line, means for calculating the inclination of the partition line candidate line at each of the representative points, and each of the parameters C, A,
Based on a first equation including B, D, and H, the image coordinates of the representative point, and the inclination of the candidate partition line, a product Z (= B * H) of the parameters B and H, and a parameter C , D
Means for estimating the parameter B; means for storing the past value of the product Z; means for obtaining the parameter B based on a second equation including the past value of the product Z; 2 or more representative points are selected for one of the line and the right partition line candidate line, and the image coordinates of each representative point, the parameters B, C, D, and the parameters C, A, B, D, Means for estimating parameters A and H based on a third equation including H, and including in the road image based on the obtained estimated values of C, A, B, D and H The feature is that a partition line position to be obtained is determined.

The invention according to claim 2 is characterized in that the first equation is represented by the above-mentioned equation (1).

The third aspect of the present invention is characterized in that the second equation is represented by the above-mentioned equation (2).

The invention according to a fourth aspect is characterized in that the third equation is expressed by the above-mentioned equation (3).

According to a fifth aspect of the present invention, there is provided means for monitoring a change in the product Z and calculating a total time Tp when the product Z vibrates an integral number of times. It is characterized in that it is set by the aforementioned equation (4).

According to a sixth aspect of the present invention, there is provided a time interval between the maximum value of the product Z and the maximum value that is an integer number of times before the maximum value,
The time between the local minimum value of the product Z and the local minimum value that is an integer number of times before that is defined as the total time Tp.

The invention according to claim 7 is characterized in that a time that is an integer number of times of the time when the product Z reaches a predetermined value is the total time Tp.

The invention according to claim 8 is characterized in that a predetermined value of the product Z is set as a current value of the product Z.

According to a ninth aspect of the present invention, the total time T
When p is short, the number of vibrations when calculating the total time Tp is increased.

[0030]

According to the first aspect of the present invention, road parameters can be accurately estimated even in the case of a one-sided white line (partition line). This means that, in addition to the information in the above equation (7),
Parameter H whose rate of change of road curvature is equivalent to camera height
This is due to the fact that it is sufficiently small compared to the rate of change. This assumption is applicable except when the road turns sharply, and is therefore sufficiently practical.
Further, it is possible to use information of a pseudo white line other than the one-sided white line (true white line), so that the robustness of estimation can be improved.

According to the second aspect of the present invention, since the above equation (1) is used as the first equation, highly accurate estimation is possible.

According to the third aspect of the present invention, since the above equation (2) is used as the second equation, highly accurate estimation is possible.

According to the fourth aspect of the present invention, since the above equation (3) is used as the third equation, highly accurate estimation is possible.

According to the fifth aspect of the present invention, it is possible to achieve both the noise removal characteristic and the followability to the true value at a higher level. This is because the assumption is established with higher accuracy in parameter estimation.

According to the invention of claim 6, it is possible to realize the function of frequency detection in claim 5 by a simple method.

According to the invention of claim 7, it is possible to realize the function of detecting the frequency in claim 5 by a simple method.

According to the invention of claim 8, it is possible to realize the function of frequency detection in claim 5 by a simple method.

According to the ninth aspect of the present invention, the first, second, and third aspects of the present invention are provided.
It is possible to more reliably realize the frequency detection function in 3, 4, and 5.

[0039]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing a vehicle 2 on which a partition line recognition device 1 according to the present invention is mounted. As shown in FIG. 1, a partition line recognition device 1 includes a monocular CCD camera (photographing means) 3 for photographing an image in front of a vehicle 2,
A processor 4 for recognizing a white line on the road in front of the vehicle based on an image captured by the CD camera 3. In the following description, a partition line drawn on a road is described as a white line, but the present invention is not limited to this, and can be similarly applied to a partition line such as a yellow line. .

FIG. 2 is a flowchart showing a procedure of image processing by the partition line recognition device 1 according to the first embodiment of the present invention. Hereinafter, the details will be described for each step.

<S1; Fetching Forward Image> The procedure is the same as that shown in the conventional example. The CCD camera 3 captures an image of the front of the vehicle at a predetermined cycle, and the processor 4 captures image information at a predetermined cycle.

<S2; Window Setting> The procedure is the same as that of the conventional example, and the center coordinate xnc of the window in the x direction is calculated based on the road model estimated at the time of the previous image capturing. The width in the x direction is set in consideration of the movement of the road white line on the image. The width in the y direction is fixed.

<S3; White Line Candidate Point Selection> This is the same as the procedure shown in the conventional example (the processing of step s103 in FIG. 8).

<S4: White Line Candidate Line Detection> The procedure is the same as that shown in the conventional example, and a line of white line candidate points in the window is linearly approximated by Hough transform. A straight line passing through the coordinates {x, y} can be expressed by the following equation (8) using the parameters a and b.

X = ay + b (8) Assuming that the coordinates of the white line candidate point are {x, y}, the parameter a
Is determined, the parameter b can be calculated by the equation (8).

Thus, an arrangement as shown in FIG. 3 can be obtained. A blank in the array means zero. The true value is included in the combination of {a, b} where 1 stands. This operation is performed for all the white line candidate points. As a result, the arrangement shown in FIG. 4 can be obtained.

The number zr of the array elements {ar, br} represents the number of white line candidate points through which the straight line given by the following equation (9) passes.

X = ar * y + br (9) An array element whose number of array elements is equal to or more than a predetermined value is set as a white line candidate point.
The candidate white line in the N-th window is stored in the following form.

{An1, bn1), {an2, bn2},...} For example, in the case of a double white line, an1 and an2 are the same.

<S5: Calculation of Representative Point Coordinate> The representative point is an intersection between the white line candidate line and the upper side of the window. Assuming that the y-coordinate of the upper side of the N-th window is yn, the x-coordinate xnm is expressed by the following equation (10).

Xnm = anm * yn + bnm (10) The suffix n of x represents the window number, the suffix m represents the white line candidate line number in the window, and the suffix of y represents the window number. The set of image coordinates of the representative points is as follows.

{X11, y1},..., {Xnm, yn},
··｝ <s6: Candidate line inclination calculation> m in the n-th window
The inclination of the white line candidate line on the image is obtained by the following equation (11).

[0053]

(Equation 7) <S7; Estimation of B * H, C, D> In many cases, white line candidate lines including pseudo white lines can be considered to be parallel to the true white line, and it can be said that equation (12) holds.

[0054]

(Equation 8) Rewriting the m-th candidate in the n-th window,
Equation (13) is obtained.

[0055]

(Equation 9) In the equation (13), all of the parameters other than the road parameters B, C, D, and H are known. A method of estimating parameters B, C, D, and H by simultaneously formulating equation (13) over a plurality of representative points includes a least square method and a Kalman filter, but details thereof are omitted here. However, it should be noted that B and H can be estimated only by their products due to the formula structure. As a result, B * H,
C and D are estimated. C and D are the current estimated values.

<S8; Creation of GAH> From now, (Nv
-1) Product B * up to before sample (Nv is the number of samples)
A set of H (= Z) is defined as GBH, a memory for storing the set GBH is prepared, and updated at each sampling. That is, the set GBH can be represented as follows.

GBH: {Z0, Z1,... Z (Nv-1)} Here, Zn is defined by the following equation (14). The subscript n means that the data is n samples before.

Zn = Bn * Hn (14) The setting of the number of samples Nv will be described later (described in step s9).

<S9; Estimation of B> The estimated value of the parameter B is calculated by the following equation (15).

[0060]

(Equation 10) However, Hs is the CCD camera 3 assumed in the standard condition of the vehicle.
Height.

Hereinafter, the reason why the parameter B can be estimated by the equation (15) and the high estimation accuracy will be described. By transforming the numerator on the right side of the equation (15), it can be expressed as the following equation (16).

[0062]

[Equation 11] Here, the point to which attention is paid is that there is a large difference between the change rates of the parameters B and H. The parameter B (corresponding to the curvature) while the vehicle bounces (amplitude up and down) and the parameter H (corresponding to the height of the CCD camera 3) changes by one cycle.
It is reasonable to assume that does not change. Even if this is not strictly defined as one cycle, the time during which the parameter B can be regarded as constant is much larger than the cycle of the parameter H, which is satisfied by a road such as an expressway where the curvature does not change rapidly. It can be said that.

If the number of samples Nv is made large within a range in which the change in the parameter B can be considered to be small, the parameter Hn in the equation (16) approaches the average parameter Hs, and therefore, as shown in the following equation (17), B * The sum of H can be approximated.

.. + BNv-1 · HNv-1 = B · (H0 + H1 +... + HNv-1) = B · Nv · Hs (17) And both sides of the equation (17) Is divided by “Nv · Hs”, the above-mentioned equation (15) can be obtained. Therefore, the parameter B can be estimated with high estimation accuracy by the equation (15).

<S10; Estimation of A and H> A plurality of white line candidate line representative points are selected using only the left white line or only the right white line. The image coordinates of these representative points (n) are {xn, yn}.

Since the parameters B, C and D have already been determined, the simultaneous equations are substituted by substituting the parameters B, C and D and the image coordinates {xn, yn} into (18) shown below. Then, parameters A and H are estimated by applying a technique such as the least square method or the Kalman filter.

[0067]

(Equation 12) Note that the least square method and the Kalman filter are well-known technologies, and thus description thereof will be omitted.

Obviously, if the white lines on both sides can be detected, the parameters A and H may be estimated by using the conventional method.

Also, for example, even when both white lines can be detected, if one of the white lines has a small number of detection points, the present method may be used. This is because, when the number of detected white lines is small and a pseudo white line is included, a large estimation error occurs in the conventional method.

Thus, even when only one of the left white line and the right white line appears in the image captured by the CCD camera 3, the parameters B, C, D, A, and H can be accurately determined. It can be estimated.

As described above, in the partition line recognition device 1 according to the present embodiment, the change rate of the road curvature (corresponding to the parameter B) is sufficiently larger than the change rate of the parameter H corresponding to the height of the CCD camera 3. By using the small white line, only one of the left white line and the right white line is detected, and the road white line ahead of the vehicle can be recognized.

Next, a second embodiment of the present invention will be described. The hardware configuration of the partition line recognition device according to the second embodiment is the same as that of the device shown in FIG.

Hereinafter, the operation of the partition line recognition device according to the second embodiment will be described with reference to the flowchart shown in FIG. In the flowchart, step s
The processing up to 28 is the same as the processing up to step s8 shown in FIG. 2, and the processing at steps s30 and s31 is the same as the processing at steps s9 and s10 shown in FIG. That is, when the second embodiment is compared with the first embodiment, only the processing in step s29 is different, and the processing in step s29 will be described below.

<S29: Estimation of fluctuation cycle of B * H> In the processing procedure according to the flowchart shown in FIG. For this purpose, counters CA and CB are prepared.

First, the counter CA is set to "0" (step s41), and the counter CB is set to "-1" (step s42). Next, the counter CA is incremented (step s43). At this time, it is determined whether or not the maximum value of the parameter Z is detected (step s4).
4). If the local maximum is not detected (NO in step s44), the process of step s43 is repeated.

If the maximum of the parameter Z is detected (YES in step s44), it is determined whether or not the negation of the counter CB (CB! Indicates negation of CB (NOT)) is equal to -1. Is performed (step s45). If not equal (NO in step s45), both the counter CA and the counter CB are set to “0”, and the processing from step s43 is repeated.

If the negative of CB (CB!) Is -1 (YES in step s45), the value of the counter CB is incremented (step s46). Next, it is determined whether or not the negative of CB (CB!) Is equal to the number of cycles k (step s47). If not (NO in step s47), the number of samples Nv is set to the value of the counter CA, and the counter CA, And CB are both set to “0”, and the process returns to step s43 (step s49).

On the other hand, negation of CB (CB!) Is the number of cycles k
If (YES in step s47), the process returns to step s43. Thus, the number of samples Nv corresponding to k periods can be obtained. Then, the parameter B is estimated by the process of step s30 (the process of s9 shown in FIG. 2) using the number of samples Nv.

When the processing procedure according to the second embodiment is used, the approximation accuracy of the equation (17) is further improved. It has been mentioned that if the number of samples Nv is set to a small value, the average value of Hn will deviate from Hs. For example, when the parameter H changes in a sine shape, for example, the average value of the plus half cycle is the camera height Hs in the standard state.
It is clear that the position shifts to the more positive side. It is also apparent that the average value of the half cycle on the minus side shifts to the minus side of the camera height Hs. As a result, when this half cycle is used, the average value of the parameter H also changes in a sine shape.

However, the number of samples Nv must not be set excessively large. During this time, the parameter B is assumed to be constant. If the number of samples Nv is increased, the followability of the estimation of the parameter B to the true value is deteriorated. As a technique for solving this trade-off, the number of times of sampling Nv can be just set to a constant multiple of the period.
For example, when the parameter H changes in the form of “sine waveform + constant offset”, the offset in the average of one cycle is constant regardless of which time is specified. Therefore, when Tp is the total time of an integral number of vibrations and Ts is the sampling interval, it is desirable to set the number of samples Nv so that the relationship of Nv = Tp / Ts holds.

The second embodiment is one of the realization methods. Since the approximation accuracy of (17) is improved, the accuracy of the resulting estimated value of the road parameter is also improved as compared with the first embodiment.

When a change occurs, for example, when it changes in a sine shape, the time from when a certain value is taken to when the next value is taken is also one cycle. Generally, the approximation accuracy is improved although it does not have a sine shape, but when considered in several cycles. Therefore, the fluctuation period may be estimated with a time interval at which the parameter Z takes a specific value. Also,
If the specific value is used as the current value of the parameter Z, there is no estimation delay, and the estimation performance is improved.

As described above, in the partition line detecting device according to the second embodiment, a suitable number of samples Nv can be obtained, so that each parameter B,
C, D, A, H can be estimated.

Next, a third embodiment of the present invention will be described. The hardware configuration of the partition line recognition device according to the third embodiment is the same as that of the device shown in FIG.

Hereinafter, the operation of the partition line recognition device according to the third embodiment will be described.

The third embodiment is similar to the second embodiment shown in FIG.
Step s of the flowchart according to the operation of the embodiment
Steps other than step 29 are the same, and the processing operation of step s29 'will be described below.

<S29 '; Estimation of fluctuation period of B * H> FIG.
Is a flowchart showing the procedure for estimating the variation cycle of B * H (= Z). In the processing procedure shown in the flowchart, the number of samples Nv corresponding to the fluctuating k cycle is detected. For this purpose, counters CA and CB are prepared.

First, the counter CA is set to "0" (step s51), and the counter CB is set to "-1" (step s52). Next, the counter CA is incremented (step s53). At this time, it is determined whether or not the maximum value of the parameter Z is detected (step s5).
4). If the local maximum is not detected (NO in step s54), the process of step s53 is repeated.

When the maximum of the parameter Z is detected (YES in step s54), it is determined whether or not the negative value (CB!) Of the counter CB is equal to -1 (step s55). If they are not equal (NO in step s55), both the counter CA and the counter CB are set to “0”, and the processing from step s53 is repeated (step s58).

If the negative of CB (CB!) Is -1 (YES in step s55), the value of counter CB is incremented (step s56). Next, it is determined whether or not the negative of CB (CB!) Is equal to the number of cycles k (step s57). If not (NO in step s57), it is determined whether or not the number of samples Nv is larger than the counter CA. Is performed (step s59). If the number of samples Nv is smaller than the counter CA (NO in step s59), the number of samples N
v is set to the value of the counter CA, and both the counters CA and CB are set to “0”, and the processing from (step s53) is repeated.

On the other hand, the counter C
If it is larger than A (YES in step s59),
The value of the cycle number k is incremented (step s6)
0), and repeat the processing from step s53 again. Thus, the number of samples Nv corresponding to k periods can be obtained. Then, the parameter B is estimated by performing the process of step S30 (the process of step s9 shown in FIG. 2) using the number of samples Nv.

The third embodiment differs from the above-described second embodiment in that when the vibration time of a predetermined number of times is shorter than a predetermined time, the number is incremented. In a normal bounce, the maximum is 1 in one vibration
For example, in the case where there is a thrust due to irregular road surface, the parameter H between the maximum and the maximum is much larger than the camera height Hs at the time of the standard posture of the vehicle (H between the maximum and the maximum at the time of the standard posture of the vehicle). Hs is always smaller).

It is clear that the average in that time is equal to or higher than the camera height Hs and a parameter estimation error occurs (the average in that time is equal to or less than Hs, and the parameter estimation error occurs. Is).

In this case, attention has been paid to the fact that the time interval between the local maximums and the local time between the local minimums is small. That is,
It is assumed that the above phenomenon occurs when the vibration time is an integral number of times, and the estimation error can be reduced by increasing the integral number to a necessary minimum. That is, when the total time Tp of the integral number of vibrations is short, the number of vibrations for calculating the total time Tp may be increased.

As described above, in the partition line detecting device according to the third embodiment, a suitable number of samples Nv can be obtained, so that each parameter B,
C, D, A, H can be estimated.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a partition line recognition device according to each embodiment of the present invention.

FIG. 2 is a flowchart illustrating a processing procedure of the partition line recognition device according to the first embodiment of the present invention.

FIG. 3 is a first explanatory diagram showing a Hough transform.

FIG. 4 is a second explanatory diagram showing the Hough transform.

FIG. 5 is a flowchart illustrating a processing procedure of the partition line recognition device according to the second embodiment and the third embodiment of the present invention.

FIG. 6 is a flowchart illustrating a processing procedure for obtaining a number of samples Nv according to the second embodiment.

FIG. 7 is a flowchart illustrating a processing procedure for obtaining the number of times of sampling Nv according to the third embodiment.

FIG. 8 is a flowchart showing a processing procedure of a conventional partition line recognition device.

FIG. 9 is an explanatory diagram showing a window set along a white line.

FIG. 10 is an explanatory diagram showing white line candidate points;

FIG. 11 is an explanatory diagram showing white line candidate lines obtained by Hough transform.

FIG. 12 is an explanatory diagram showing representative points of white line candidate lines.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Partition line recognition apparatus 2 Vehicle 3 CCD camera 4 Processor

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference) 2F065 AA11 BB12 CC40 FF04 FF26 GG10 JJ03 JJ26 PP01 QQ01 QQ17 QQ18 QQ29 QQ51 UU05 5B057 AA16 BA11 CA13 CB12 CC03 CE15 CH01 DA07 DB03 DC08 DC32 5H180 AA01 CC02 CC02 CC02 CC24 CE02 CA14 DA01 FA03 FA33 FA67 FA69 FA76 GA17

Claims (9)

[Claims] 1. A monocular photographing means for photographing a road image in front of a vehicle, wherein a parameter C corresponding to a yaw angle and a parameter corresponding to a lateral displacement are discretely based on an image photographed by the photographing means. A, a parameter B corresponding to a curvature, a parameter D corresponding to a pitch angle, and a parameter H corresponding to a camera height, and a road partition for recognizing a partition line included in the road image based on the estimated parameters. In the line recognition device, from a video image captured by the image capturing unit, a partition line candidate line serving as a partition line candidate is detected, and image coordinate of a representative point set on the partition line candidate line is calculated. Means for determining the inclination of the partition line candidate line at each of the representative points; a first equation including each of the parameters C, A, B, D, and H; the image coordinates of the representative point; Based on the slope of Therefore, the product Z (= B *) of the parameters B and H
H) means for estimating parameters C and D; means for storing past values of the product Z; means for obtaining the parameter B based on a second equation including past values of the product Z And selecting one or more of the representative points for either the left partition line candidate line or the right partition line candidate line, and selecting the image coordinates of each representative point, the parameters B, C, D, and the parameter. Means for estimating parameters A, H based on a third equation including C, A, B, D, H, wherein the estimated values of C, A, B, D, H A partition line recognition device for a road, wherein a partition line position included in the road image is determined based on the road image. 2. The road partitioning line recognition device according to claim 1, wherein the first equation is expressed by the following equation (1). (Equation 1) Here, fh and fv are constants in the horizontal and vertical directions when the three-dimensional coordinates in the real space are perspective-transformed into two-dimensional image coordinates, n is the number of each representative point, and n0 is the total number of representative points. 3. The road partition line recognition device according to claim 1, wherein the second equation is expressed by the following equation (2). (Equation 2) Here, Hs is a parameter H assumed in a vehicle standard state,
Nv is the number of past values of the product Z. 4. The road partition line recognition device according to claim 1, wherein the third equation is expressed by the following equation (3). (Equation 3) Here, i is 0 for the left white line, 1 for the right white line, E0 is the width of the left and right white lines, n is the number assigned to each representative point, and nf is the total number of representative points obtained by summing the left and right. 5. A monitor for monitoring a change in the product Z and calculating a total time Tp when the product Z vibrates an integer number of times.
The road partition line recognition device according to any one of claims 3 and 4, wherein the number Nv of past values is set by the following equation (4). Nv = Tp / Ts (4) where Ts is a sampling interval. 6. A time interval between a local maximum value of the product Z and a local maximum value thereof an integral number of times before, or a time interval between a local minimum value of the product Z and a local minimum value thereof an integral number of times before the total. Time T
The road partition line recognition device according to claim 5, wherein p is set to p. 7. The road partitioning line recognition device according to claim 5, wherein a period of time at which the product Z takes a predetermined value becomes an interval of an integer number of times is the total time Tp. 8. The road partition line recognition device according to claim 7, wherein the predetermined value of the product Z is a current value of the product Z. 9. An apparatus for recognizing a partition line for a road, wherein when the total time Tp is short, the number of vibrations for calculating the total time Tp is increased.