JP2005056322A - White line estimation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent no detection of white line candidate points when the luminance of an input image is inclined to saturation and to improve the detection performance of a white line. <P>SOLUTION: In detection of white line candidate points, whether either one of the positive maximum point of an edge and the negative minimum point of the edge is to be used or both of them are to be used is determined in accordance with a camera contour correction system. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a white line estimation device that estimates a white line from an image captured by a camera mounted on a vehicle traveling on a road.

As a prior art of the present invention, a white line estimation device for estimating a white line on a road used in a lane departure warning device as a driving support device will be described. Patent documents 1, 2, and 3 are mentioned as a prior art example.
The lane departure warning device includes an image capturing unit such as a camera, an image processing unit that processes the captured image, and an alarm generation unit.

The image processing unit detects the left and right white lines from the image obtained from the image capturing unit, and estimates the lane in which the host vehicle is traveling.
The warning generator calculates the left and right distances from the vehicle at the lane boundary, and if either is below the threshold, that is, if the vehicle is closer to the lane boundary than the distance set by the threshold, the driver Give an alarm.
The camera has a contour correction circuit incorporated therein. The contour correction circuit is a circuit that detects a point having a large luminance change in the image output from the light receiving element and performs processing for emphasizing the point. FIG. 1 shows a specific example of contour correction. Assuming that FIG. 1A is an output image of the light receiving element, the contour correction circuit can obtain images such as those shown in FIGS. 1B, 1C, and 1D.

  FIG. 1B is an output example of a contour correction circuit that enhances both rising and falling edges when an image is scanned horizontally from left to right.

  Hereinafter, the term “rising edge” or “falling edge” simply represents an edge when the image is scanned horizontally from left to right.

  FIG. 1C shows an output example of the contour correction circuit that emphasizes only the falling edge.

  FIG. 1D shows an output example of the contour correction circuit that emphasizes only the rising edge.

  The image processing unit includes a white line detection device.

  In such a conventional white line detection circuit, when the contour correction is a method of emphasizing only one of the rising edge and the falling edge, or when neither is emphasizing, the brightness of the road surface is increased. When it is higher, the white line cannot be detected stably.

  FIGS. 2A and 2B show an example in which the road surface has high luminance.

FIG. 2A shows the output of the light receiving element, and FIG. 2B shows the output of the contour correction circuit that emphasizes only the falling edge. This is a situation where the sunlight in the daytime is strong. The brightness of the road surface becomes high, and there is almost no difference in brightness between the white line and the road surface. From FIG. 2B, only the falling edge is stably detected. The conventional method is based on the premise that both rising and falling edges are detected stably, so that the white line is not detected stably.
JP-A-10-214326 JP 2002-175535 A JP 2002-175534 A

  An object of the present invention is to provide a white line estimation device capable of improving the performance of white line detection processing.

  The white line estimation device according to the present invention includes an imaging unit that captures an image, and an output unit that outputs an image signal in which at least a rising edge or a falling edge is emphasized to correct a contour when scanning a captured image by the imaging unit. And setting means for setting whether the contour correction by the output means is performed by emphasizing the rising edge, emphasizing the falling edge, emphasizing both the rising edge and the falling edge, and output by the output means A first detection means for detecting an edge image based on the image signal, and a second detection for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means. If the rising edge enhancement is set by the means and the setting means, the white line candidate point is detected by the positive maximum value obtained by the second detecting means. If the falling edge enhancement is set by the setting means, the white line candidate point is detected by the negative minimum value obtained by the second detection means, and the rising edge and the falling edge are detected by the setting means. When both enhancements are set, a third detection unit that detects a white line candidate point based on the positive maximum value and the negative minimum value obtained by the second detection unit, and the third detection unit By detecting a straight line from the detected white line candidate points, a fourth detection means for detecting a white line candidate and a certainty factor of the white line candidate detected by the fourth detection means are obtained, and the certainty factor of the white line candidate is predetermined. And an estimation means for estimating the white line candidate as a white line when the value exceeds the value.

  The white line estimation device according to the present invention is mounted on a vehicle and has an imaging unit that captures an image, and an image signal that is contour-corrected by emphasizing at least a rising edge or a falling edge when scanning an image captured by the imaging unit. Output means, and setting means for setting whether the edge correction by the output means is performed by enhancement of rising edge, enhancement of falling edge, enhancement of both rising edge and falling edge, and Based on the image signal output from the output means, a first detection means for detecting an edge image, and a positive maximum value and a negative minimum value are detected from each edge value of the edge image detected by the first detection means. For the second detection means that performs the detection and the area close to the host vehicle, the white line candidate points are detected by the positive maximum value and the negative minimum value obtained by the second detection means. In the case where the rising edge emphasis is set by the setting means, the white line candidate point is detected by the positive maximum value obtained by the second detection means for the area separated from the own vehicle. When emphasis of the falling edge is set by the setting means, the white line candidate point is detected by the negative minimum value obtained by the second detection means, and both the rising edge and the falling edge are detected by the setting means. Is set by the third detection means for detecting the white line candidate point based on the positive maximum value and the negative minimum value obtained by the second detection means, and the third detection means. By detecting a straight line from the white line candidate points, the fourth detection means for detecting the white line candidate and the confidence of the white line candidate detected by the fourth detection means are obtained, and the confidence of the white line candidate During more than a predetermined value, and a estimation means for estimating the white line candidate white line and.

  The white line estimation device according to the present invention is mounted on a vehicle traveling on a road having a white line, and has at least a rising edge or a falling edge when scanning an image picked up by the image pickup means and the image pickup means. An output unit that outputs an image signal that has been enhanced and contour corrected, a first detection unit that detects an edge image based on the image signal output from the output unit, and an edge image detected by the first detection unit Second detection means for detecting a positive maximum value and a negative minimum value from each edge value, and a white line candidate point is detected by a positive maximum value and a negative minimum value obtained by the second detection means. 3 detecting means, a fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means, and a detecting means by the fourth detecting means. When the certainty factor of the white line candidate is determined and the certainty factor of the white line candidate is equal to or greater than a predetermined value, the white line candidate is estimated as a white line, and the contour correction is not estimated by the second estimating unit. When the brightness of the road surface of the image captured by the image capturing means tends to be saturated, the ineffective means for invalidating the estimation of the white line by the first estimating means and the vehicle traveling on the road having the white line are mounted, An image pickup means for picking up an image, an output means for outputting an image signal whose contour is corrected by emphasizing at least the rising edge or the falling edge when scanning a captured image by the image pickup means, and an image output by the output means First detection means for detecting an edge image by a signal, and first detection means for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means. Detection means, third detection means for detecting white line candidate points based on the positive maximum value and negative minimum value obtained by the second detection means, and white line candidate points detected by the third detection means By detecting a straight line from the fourth detection means for detecting a white line candidate and the confidence of the white line candidate detected by the fourth detection means, the confidence of the white line candidate is equal to or greater than a predetermined value. The first correction means for estimating the white line candidate as a white line, and the contour correction by the output means based on the luminance around the white line candidate detected by the fourth detection means, the rising edge enhancement and the falling edge enhancement. A second estimation unit that estimates whether the rising edge or the falling edge is emphasized, and the second estimation unit does not estimate the contour correction, and the road of the captured image by the imaging unit Plane Invalidity means for invalidating white line estimation by the first estimation means when the luminance tends to be saturated.

  The white line estimation device according to the present invention includes an imaging unit that captures an image, and an output unit that outputs an image signal in which at least a rising edge or a falling edge is emphasized to correct a contour when scanning a captured image by the imaging unit. A first detection unit that detects an edge image based on an image signal output from the output unit, and a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection unit. The second detection means for detecting the white line candidate point by the positive maximum value and the negative minimum value obtained by the second detection means, and the third detection means The fourth detection means for detecting a straight line from the detected white line candidate points and the confidence of the white line candidate detected by the fourth detection means are obtained, and when the confidence of the white line candidate is equal to or greater than a predetermined value, Contour correction by the output means from the first estimation means for estimating the line candidate as a white line, and the luminance around the white line candidate detected by the fourth detection means, emphasizes the rising edge, emphasizes the falling edge, and rises Second estimation means for estimating whether the enhancement is performed by emphasizing both the edge and the falling edge.

  The white line estimation device according to the present invention is mounted on a vehicle traveling on a road having a white line, and has at least a rising edge or a falling edge when scanning an image picked up by the image pickup means and the image pickup means. An output unit that outputs an image signal that has been enhanced and contour corrected, a first detection unit that detects an edge image based on the image signal output from the output unit, and an edge image detected by the first detection unit Second detection means for detecting a positive maximum value and a negative minimum value from each edge value, and a white line candidate point is detected by a positive maximum value and a negative minimum value obtained by the second detection means. 3 detecting means, a fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means, and a detecting means by the fourth detecting means. When the certainty level of the white line candidate is determined and the certainty level of the white line candidate is equal to or greater than a predetermined value, the first estimation unit that estimates the white line candidate as a white line and the white line candidate detected by the fourth detection unit Second estimation means for estimating whether the contour correction by the output means is performed from surrounding luminance by enhancement of rising edge, enhancement of falling edge, enhancement of both rising edge and falling edge; Invalidity means for invalidating white line estimation by the first estimation means when contour correction is not estimated by the second estimation means, and the road surface brightness of the image captured by the imaging means tends to be saturated; Have

  As described above in detail, according to the present invention, it is possible to provide a white line estimation device capable of improving the performance of white line detection processing.

A lane departure warning device according to an embodiment of the present invention will be described below with reference to the drawings.
As shown in FIG. 3, the lane departure warning device includes an image capturing unit 1 such as a camera, an image processing unit 2 that processes the captured image, a user interface 3 in which various settings are made, and an alarm generating unit 4. Has been. For example, doze detection is performed by white line detection. In particular, the white line detection can be performed while avoiding the drawbacks of the white line detection that is retrofitted using a dedicated camera (wide angle) for parking.

  That is, as the image photographing unit 1, a camera attached so as to photograph the front or rear of the host vehicle is used like the camera A or the camera B in FIG. An image as shown in FIG. 1A is obtained from such a camera. FIG. 1A shows an example in which the left and right white lines of the own lane exist in the input image.

  As shown in FIG. 5, the image photographing unit 1 includes a lens 11, a light receiving element 12, and a contour correction circuit 13.

  The contour correction circuit 13 is a circuit that detects a point with a large luminance change in the image output from the light receiving element 12 and performs a process of emphasizing the point. FIG. 1 shows a specific example of contour correction. When FIG. 1A is an output image of the light receiving element, the contour correction circuit 13 obtains images such as those shown in FIGS. 1B, 1C, and 1D. The contour correction circuit 13 differs depending on the model number and manufacturer of the camera, and is selected depending on the cost.

  FIG. 1 (b) shows a rising edge when the image is scanned horizontally from left to right (where the luminance changes from a dark place on a road to a bright place on a white line: dark to bright) and a falling edge ( This is an output example of the contour correction circuit 13 that emphasizes both the brightness of the bright white line and the brightness of the dark road (bright to dark).

  Hereinafter, the term “rising edge” or “falling edge” simply represents an edge when the image is scanned horizontally from left to right.

  FIG. 1C shows an output example of the contour correction circuit 13 that emphasizes only the falling edge when the image is scanned horizontally from left to right.

  FIG. 1D shows an output example of the contour correction circuit 13 that emphasizes only rising edges when the image is scanned horizontally from left to right.

The image processing unit 2 detects white lines on the left and right sides based on the image obtained from the image capturing unit 1 and the setting contents of the user interface 3, and estimates the lane in which the host vehicle is traveling. Details of this processing will be described later. In the image processing unit 2, a parameter storage unit 2a is set, and various parameters are set. For example, contour correction is performed by the user interface 3 by the contour correction circuit 13 by emphasizing only the falling edge, emphasizing only the rising edge, and emphasizing both the falling edge and the rising edge, or the falling edge and the rising edge. Either one of those not emphasizing is set, and this setting content is stored in the parameter storage unit 2a,
The alarm generation unit 4 calculates the distance DL and the distance DR from the own vehicle with respect to the lane boundary as shown in FIG. A warning is given to the driver when approaching distance Tdst.

  A processing flow of the alarm generation unit 4 is shown in FIG. The calculation of the distance DL and the distance DR is specifically performed as follows. The estimated left and right lane boundaries are converted from the xy coordinate system of the input image as shown in FIG. 8 to an XZ coordinate system as seen from above as shown in FIG. 9 (ST1).

  When the length from the camera center to the left and right vehicle ends is WIDTH_L and WIDTH_R, respectively, and in the XZ coordinate system, the distance between the Z axis and the left and right lane boundaries is DC_L and DC_R, respectively. The calculation is performed as follows: DL = DC_L-WIDTH_L, DR = DC_R-WIDTH_R (ST2). The coordinate system conversion formula can be determined from the camera parameters.

  When the distance DL and the distance DR from the own vehicle with respect to the lane boundary are calculated and one of them is equal to or smaller than the threshold Tdst, that is, when the own vehicle approaches the lane boundary than the distance Tdst set by the threshold (ST3) Then, a lane departure warning is generated (ST4), and a warning is given to the driver (driver) by a display by a display unit (not shown) or a warning by a buzzer.

  Hereinafter, the operation of the image processing unit 2 will be described in detail. FIG. 10 shows a processing flow of the image processing unit 2. In the image input process (ST11), the image from the camera 1 is A / D converted and input at a certain time interval. In the image input process, the video signal from the camera is digitized at regular time intervals and captured as an image.

  Specifically, images are captured at time t0, t0 + δt (= t1), t0 + 2δt (= t2),. Here, t0 and δt are a processing start time and a time interval for capturing an image, respectively. For example, as shown in FIG. 10, in the image input process (ST11), after capturing an image at time ti, during time δt, edge detection process (ST12), white line candidate point detection process (ST13) as preprocessing is performed. ), White line candidate detection processing (ST14), lane estimation processing (ST15), contour correction method estimation processing (ST16), estimated lane verification processing (ST17), and image input processing (ST11) again at time ti + δt = ti + 1 The process is repeated so that an image is captured. In the following, description will be given of a case where image input is performed at time ti and a digitized input image Iti (x, y) is obtained.

  In the preprocessing, processing such as noise removal such as a smoothing filter is performed on the input image.

That is, in the edge detection process as the pre-processing, the edge image Eti (x; y) is obtained from the input image by performing a filter operation as shown in FIG. 11 on the input image Iti (x; y). Expressed as a formula:
Eti (x, y) =-Iti (x-2y) -2 Iti (x-1, y)
+2 Iti (x + 1, y) + Iti (x + 2, y)
It becomes like this.

  In the white line candidate point detection processing, white line candidate points are detected from the edge image by a method corresponding to the type of contour correction of the contour correction circuit 13 stored in the parameter storage unit 2a.

  For example, when the contour correction emphasizes only the rising edge, a white line candidate point is detected from the positive maximum point of the edge.

  Further, when the contour correction emphasizes only the falling edge, a white line candidate point is detected from the negative minimum point of the edge.

  Further, when the contour correction emphasizes both the rising edge and the falling edge, a white line candidate point is detected from the set of the positive maximum point of the edge and the negative minimum point of the edge.

  The white line candidate point detection process will be described with reference to the flowchart shown in FIG.

  First, when the contour correction of the contour correction circuit 13 stored in the parameter storage unit 2a emphasizes only the rising edge (YES in ST20, YES in ST21), a white line is detected from the positive maximum point of the edge. Candidate points are detected (ST22, 23).

  Further, when the contour correction of the contour correction circuit 13 stored in the parameter storage unit 2a emphasizes only the falling edge (YES in ST20, NO in ST21, YES in ST24), the negative edge is detected. White line candidate points are detected from the local minimum points (ST25, ST26).

  When the contour correction of the contour correction circuit 13 stored in the parameter storage unit 2a emphasizes both the rising edge and the falling edge (YES in ST20, NO in ST21, NO in ST24). Alternatively, when the contour correction of the contour correction circuit 13 stored in the parameter storage unit 2a does not emphasize both the rising edge and the falling edge (NO in ST20), the positive maximum point of the edge is White line candidate points are detected from the midpoint of the negative minimum point of the edge (ST27, 28, 29).

  For example, detection of white line candidate points when the contour correction emphasizes both the rising edge and the falling edge will be described.

  That is, when both the rising edge and falling edge detection is performed on the input image of FIG. 13, an edge image as shown in FIG. 14A is obtained. In FIG. 14A, the point where the absolute value of the edge value is large is represented in black.

  FIG. 14B shows the edge values along the straight line L in FIG. In FIG. 14 (b), it has a positive maximum value at the point where the road surface changes from white to left as shown by A and A '. Further, it has a negative minimum value at a point where the white line changes from the left to the right as in B and B 'and the road surface. Therefore, the edge image Eti (x, y) is scanned horizontally from left to right, and the left and right are a set of positive maximum values and negative minimum values, respectively, and the distances in the x direction are white lines. An appropriate width is detected, and the middle point is set as a white line candidate point. The points on the straight line L in FIG. 14A are selected as a set of positive maximum values and negative minimum values of A and B, and A 'and B', and the midpoint is a white line candidate point.

In the white line candidate detection process, a straight line is detected from the white line candidate points and set as a white line candidate. For this, for example, Hough transform is used. First, voting is performed on the parameter plane for each white line candidate point. One point on the parameter plane corresponds to one straight line on the input image plane. Specifically, one point (θ0, ρ0) on the ρθ parameter plane in FIG. 15 corresponds to a straight line as shown in FIG. The straight line in FIG. 16 is a straight line in which the length of the perpendicular from the origin to the straight line is ρ0 and the angle with the x axis is θ0. The vote for the white line candidate point P (xp, yp) on the xy plane is the curve on the ρθ parameter plane.
ρ = xp · cos θ + yp · sin θ
Follow along.

Next, after voting for each white line candidate point, a peak on the ρθ parameter plane is detected. A straight line on the xy plane corresponding to each peak is obtained and set as a white line candidate.
FIG. 18 shows the result of detecting white line candidates from the white line candidate points in FIG. White line candidates such as LC1 and LC2 are detected.

  In the lane estimation process, the lane in which the vehicle is traveling is estimated from the detected white line candidates. The estimated lane obtained by this processing is represented as a set of left and right straight lines in the input image. First, a set of left and right white line candidates is selected from white line candidates obtained by white line candidate detection. When there are a plurality of white line candidate groups, the certainty factor is calculated for the white line candidate group, and the white line candidate group having the highest certainty factor is selected. As a simple example of the certainty factor of a set of white line candidates, the number of white line candidate points on the white line candidates is considered.

  In addition, the certainty factor may be calculated from the intensity of the maximum and minimum values of the edge image used when detecting the white line candidate point, the brightness contrast around the white line candidate, and the like.

  From FIG. 18, LC1 and LC2 are selected as a pair of left and right white lines. Next, the lane at time ti is estimated from the set of white line candidates obtained. As a simple method, it is conceivable to use the selected white line candidate as it is as the estimated lane at time ti. Moreover, you may estimate using a Kalman filter.

  FIG. 19 shows the flow of the contour correction method estimation process.

  The contour correction method is estimated from the luminance around the positive maximum point of the edge where the white line candidate is detected and the negative minimum point.

  When the contour correction method has been estimated (YES in ST31) and when the luminance of the input image tends to be saturated (YES in ST32), the contour correction method is not estimated. This is because the luminance of the input image is This is because if there is a saturation tendency, the white line candidate is not detected correctly because the luminance difference between the white line and the road surface disappears, and as a result, the contour correction method may not be estimated correctly. . Whether or not the luminance of the input image tends to be saturated is calculated as an average luminance value for a fixed region corresponding to the road surface of the input image, and is determined to be saturated when it is larger than the threshold value Tl.

  When the contour correction circuit 13 emphasizes both rising and falling edges, highlighting only the rising edge, highlighting only the falling edge, and not emphasizing both the rising and falling edges, the white line periphery 20, FIG. 21, FIG. 22, and FIG. 23, respectively, show the luminances of. 20A, FIG. 21A, FIG. 22A, and FIG. 23A all show the output of the contour correction circuit 13. FIG. 20 (a), FIG. 21 (a), FIG. 22 (a), and FIG. 23 (a) represent the luminance around the white line, respectively, and FIG. 20 (b), FIG. 21 (b), and FIG. 22 (b) and FIG. 23 (b).

With respect to the positive maximum point of each edge where the white line candidate is detected, as shown in FIG. 20B, a point A that is separated from the maximum point by a certain distance Dx in the negative direction of x is taken,
Iti (A) -Iti (positive maximum point of the edge)> Tdiff (1)
The number of local maximum points that satisfy the above is counted (ST33). Here, Iti (A) represents the luminance value of the image at point A. Tdiff is a positive threshold value. If the ratio of the positive maximum points of the edge where the white line candidate is detected satisfies the condition Tpcnt or more (YES in ST34), it is determined that the rising edge is emphasized by the contour correction circuit 13 (ST35). . If the ratio of the positive maximum points of the edge where the white line candidate is detected satisfies the condition Tpcnt or less (NO in ST34), the contour correction circuit 13 determines that the rising edge is not emphasized (ST36). The count is performed using an internal memory of the image processing circuit.

For the negative minimum point of each edge where the white line candidate is detected, as shown in FIG. 20B, a point B separated from the minimum point by a certain distance Dx in the positive direction of x is taken Iti (B) −Iti ( Edge negative minimum)> Tdiff (2)
The number of minimum points that satisfy the above is counted (ST37). If the ratio of the negative minimum points of the edge where the white line candidate is detected satisfies the condition (YES in ST38), it is determined that the falling edge is emphasized by the contour correction circuit 13 (ST39). ). If the ratio of the negative minimum points of the edge where the white line candidate is detected satisfies the condition is equal to or less than Tpcnt (NO in ST38), it is determined that the falling edge is not emphasized by the contour correction circuit 13 (ST40). .

  In the case of FIG. 20B, Iti (positive maximum point of the edge) is smaller than Iti (A). Further, Iti (negative minimum point of the edge) becomes smaller than Iti (B), and it is determined that the contour correction circuit 13 emphasizes both the rising and falling edges.

  In the case of FIG. 21B, Iti (the positive maximum point of the edge) is smaller than Iti (A). Further, Iti (negative minimum point of the edge) and I (B) are substantially the same, and it is determined that the contour correction circuit 13 emphasizes only the rising edge.

  In the case of FIG. 22B, Iti (the positive maximum point of the edge) and I (A) are substantially the same. Further, Iti (negative minimum point of the edge) becomes smaller than Iti (B), and it is determined that the contour correction circuit 13 emphasizes both the rising and falling edges.

  In the case of FIG. 23B, I (positive maximum point of the edge) and I (A) are substantially the same. Further, I (negative minimum point of the edge) and I (B) become substantially the same, and it is determined that the contour correction circuit 13 does not emphasize both the rising and falling edges.

  The flow of verification of the estimated lane is shown in FIG.

  If the contour correction method is unknown (NO in ST41) and the road surface luminance of the input image tends to be saturated (YES in ST42), it is not known whether the edge of the white line is stably detected. The estimated lane is determined to be invalid (ST43). Further, when the rising and falling edges are not emphasized by the contour correcting circuit 13 and the brightness of the road surface of the input image tends to be saturated (ST44), the white line edge is not detected stably. The estimated lane is determined to be invalid (ST43).

  Except for the above cases, the estimated lane is valid.

  That is, if the contour correction method is unknown (NO in ST41) and the road surface brightness of the input image does not tend to be saturated (NO in ST42), it is assumed that the edge of the white line is detected stably. The lane is determined to be valid (ST45). Whether the rising and falling edges are emphasized by the contour correction circuit 13 or whether one of the rising and falling edges is emphasized (ST46), the luminance of the road surface of the input image tends to be saturated. If not (NO in ST44), it is determined that the edge of the white line is stably detected, and the estimated lane is determined to be valid (ST45).

  FIGS. 25 to 28 show examples in which the estimated lane is validated and invalidated.

  FIG. 25 is an input image when the estimated lane is validated, and FIG. 26 is the luminance distribution of FIG. In the contour correction circuit, both rising and falling edges are not emphasized, but it can be seen that the luminance distribution on the road surface and the luminance distribution on the white line are sufficiently separated. In such a case, the white line is detected stably. In verification of the estimated lane, the contour correction circuit does not emphasize both rising and falling edges, but it is determined from the luminance distribution that the luminance of the road surface of the input image does not tend to be saturated, and the estimated lane in this case is valid. It is determined that there is.

  FIG. 27 is an input image when the estimated lane is validated, and FIG. 28 is the luminance distribution of FIG. The contour correction circuit 13 shows that the rising edge and the falling edge are not emphasized, and the luminance distribution on the road surface and the luminance distribution on the white line are very close. In such a case, the white line is not detected stably. In the verification of the estimated lane, the contour correction circuit 13 determines that the rising and falling edges are not emphasized and that the luminance of the road surface of the input image tends to be saturated from the luminance distribution. In this case, the estimated lane is Is determined to be invalid.

  If it is determined that the estimated lane is invalid, the alarm generation unit 4 stops the alarm generation process and notifies the driver that the lane departure warning system is not operating normally.

  As described above, in the white line candidate point detection, it is determined whether to use the positive maximum point of the edge, the negative minimum point of the edge, or both, according to the contour correction method.

  As a result, when the luminance of the input image tends to be saturated, it is possible to prevent white line candidate points from being detected, and as a result, white line detection performance is improved.

  Further, it is determined whether the estimated lane is valid or invalid from the contour correction method and the luminance distribution of the input image, and corresponds to verification of the estimated lane.

  As a result, a situation in which there is a high possibility of erroneous detection can be detected. In this case, the result of the white line detection process is determined to be invalid. As a result, the performance of the white line detection process is improved.

  Also, the contour correction method is automatically estimated from the brightness around the white line candidate.

  Thereby, adjustment of the white line detection process according to a camera can be automated.

  That is, when the lane departure system is mounted on various types of vehicles, an existing back monitor camera is often used as the camera. In such a situation, various types of back monitor cameras are assumed. Accordingly, various types of camera contour correction methods are assumed, and it is necessary to adjust the white line detection process in accordance with each method. On the other hand, the contour correction method of the camera is automatically detected, and the white line detection process can be automatically adjusted without human intervention.

  As another example, when white line candidate point detection processing is performed, different processing may be performed for a region close to the host vehicle and a region far from the vehicle.

  As shown in FIG. 29, it is known that on the road surface in the input image, the brightness tends to increase as the vanishing point is approached. This tendency is disclosed in Japanese Laid-Open Patent Publication No. 2001-273504 (in consideration of the fact that the brightness value of the road surface increases as it approaches the vanishing point, the road area is detected from the brightness information. Is not relevant.) Therefore, in the region close to the vehicle as surrounded by the thick line in FIG. 30, the tendency of luminance saturation is weak, and both rising and falling edges are detected stably. White line candidate points are detected from a set of local minimum points. In other areas, detection is performed according to the processing method of the contour correction circuit. In FIG. 31, step 50 for determining whether the point Pi is a point in the region close to the own vehicle is inserted between ST20 and ST21 in FIG. 12, and if YES, the process proceeds to step 28. If NO, Proceed to step 21. Other operations are the same.

  As a result, in detecting the white line candidate point, it is determined whether to use the positive maximum point of the edge, the negative minimum point of the edge, or both depending on whether or not the region is close to the own vehicle in the input image.

  As a result, at the position where the edge is stably detected in the input image, the white line candidate point is detected by using both the rising and falling edges. Performance is improved. When detecting the white line candidate points, when the rising and falling edges are stable, the use of both of them is less likely to detect edges other than the white line by chance, and the performance is improved.

The figure which shows the specific example of the output image by a contour correction circuit. The figure which shows the output image by an outline correction circuit. The block diagram which shows schematic structure of the lane departure warning apparatus concerning embodiment of this invention. The figure for demonstrating the vehicle-mounted position of a camera. The block diagram which shows schematic structure of an image imaging | photography part. The figure for demonstrating the distance from the own vehicle with respect to the lane boundary calculated by the warning production | generation part. The figure for demonstrating the flow of a process of an alert generation part. The figure for demonstrating the xy coordinate system of an input image. The figure for demonstrating the XZ coordinate system which looked at the own vehicle from the top. The figure for demonstrating the flow of a process of an image process part. The figure which shows an example of filter calculation. The figure for demonstrating the flow of a white line candidate point detection process. The figure which shows the input image of a white line candidate point detection process. The figure which shows the example of a detection of the edge image by edge detection, and the example of a detection of a white line candidate point. The figure which shows (rho) theta parameter plane. The figure which shows a response | compatibility with xy plane and (rho) theta parameter plane. The figure which shows the example of a detection of a white line candidate point. The figure which shows the example of a detection of a white line candidate. The figure which shows the flow of an outline correction system estimation process. The figure which shows the example of an output of a contour correction circuit, and the brightness | luminance of a white line periphery. The figure which shows the example of an output of a contour correction circuit, and the brightness | luminance of a white line periphery. The figure which shows the example of an output of a contour correction circuit, and the brightness | luminance of a white line periphery. The figure which shows the example of an output of a contour correction circuit, and the brightness | luminance of a white line periphery. The figure which shows the flow of verification of an estimated lane. The figure which shows the input image in the case of validating an estimated lane. The figure which shows the luminance distribution of the input image of FIG. The figure which shows the input image in the case of validating an estimated lane. The figure which shows the luminance distribution of the input image of FIG. The figure which shows the tendency of the luminance distribution of the road surface in an input image. The figure which shows the area | region which detects a white line from the group of a rising edge and a falling edge. The figure for demonstrating the flow of a white line candidate point detection process.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image pick-up part 2 ... Image processing part 3 ... User interface 4 ... Alarm production | generation part.

Claims (5)

Imaging means for imaging video;
An output unit that outputs an image signal that has been subjected to contour correction by emphasizing at least a rising edge or a falling edge when scanning an image captured by the imaging unit;
Setting means for setting whether the contour correction by the output means is performed by emphasizing the rising edge, emphasizing the falling edge, emphasizing both the rising edge and the falling edge;
First detection means for detecting an edge image from the image signal output by the output means;
Second detection means for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means;
When the rising edge enhancement is set by the setting means, the white line candidate point is detected by the positive maximum value obtained by the second detection means, and the falling edge enhancement is set by the setting means. If the white line candidate point is detected by the negative minimum value obtained by the second detection means, and both the rising edge and the falling edge are emphasized by the setting means, the second Third detection means for detecting a white line candidate point based on a positive maximum value and a negative minimum value obtained by the detection means;
A fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means;
An estimation unit that obtains a certainty factor of the white line candidate detected by the fourth detection unit and estimates the white line candidate as a white line when the certainty factor of the white line candidate is equal to or greater than a predetermined value;
The white line estimation apparatus characterized by comprising. An imaging means mounted on the vehicle for capturing images;
An output unit that outputs an image signal that has been subjected to contour correction by emphasizing at least a rising edge or a falling edge when scanning an image captured by the imaging unit;
Setting means for setting whether the contour correction by the output means is performed by emphasizing the rising edge, emphasizing the falling edge, emphasizing both the rising edge and the falling edge;
First detection means for detecting an edge image from the image signal output by the output means;
Second detection means for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means;
For the region close to the host vehicle, white line candidate points are detected by the positive maximum value and the negative minimum value obtained by the second detection unit, and for the region separated from the host vehicle, the setting unit If the rising edge enhancement is set by the above, the white line candidate point is detected by the positive maximum value obtained by the second detection means, and the falling edge enhancement is set by the setting means, When the white line candidate point is detected based on the negative minimum value obtained by the second detection means, and both the rising edge and the falling edge are emphasized by the setting means, the second detection means Third detection means for detecting a white line candidate point from the obtained positive maximum value and negative minimum value;
A fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means;
An estimation unit that obtains a certainty factor of the white line candidate detected by the fourth detection unit and estimates the white line candidate as a white line when the certainty factor of the white line candidate is equal to or greater than a predetermined value;
The white line estimation apparatus characterized by comprising. An image pickup means mounted on a vehicle traveling on a road having a white line and picking up an image;
An output unit that outputs an image signal that has been subjected to contour correction by emphasizing at least a rising edge or a falling edge when scanning an image captured by the imaging unit;
First detection means for detecting an edge image from the image signal output by the output means;
Second detection means for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means;
Third detection means for detecting a white line candidate point from the positive maximum value and the negative minimum value obtained by the second detection means;
A fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means;
An estimation unit that obtains a certainty factor of the white line candidate detected by the fourth detection unit and estimates the white line candidate as a white line when the certainty factor of the white line candidate is equal to or greater than a predetermined value;
Invalidity means for invalidating white line estimation by the first estimation means when contour correction is not estimated by the second estimation means, and the road surface brightness of the image captured by the imaging means tends to be saturated; ,
The white line estimation apparatus characterized by comprising. Imaging means for imaging video;
An output unit that outputs an image signal that has been subjected to contour correction by emphasizing at least a rising edge or a falling edge when scanning an image captured by the imaging unit;
First detection means for detecting an edge image from the image signal output by the output means;
Second detection means for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means;
Third detection means for detecting a white line candidate point from the positive maximum value and the negative minimum value obtained by the second detection means;
A fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means;
A first estimation unit that obtains a certainty factor of the white line candidate detected by the fourth detection unit and estimates the white line candidate as a white line when the certainty factor of the white line candidate is equal to or greater than a predetermined value;
The contour correction by the output means is performed based on the luminance around the white line candidate detected by the fourth detection means by any of the rising edge enhancement, the falling edge enhancement, and both the rising edge and the falling edge enhancement. Second estimating means for estimating whether or not
The white line estimation apparatus characterized by comprising. An image pickup means mounted on a vehicle traveling on a road having a white line and picking up an image;
An output unit that outputs an image signal that has been subjected to contour correction by emphasizing at least a rising edge or a falling edge when scanning an image captured by the imaging unit;
First detection means for detecting an edge image from the image signal output by the output means;
Second detection means for detecting a positive maximum value and a negative minimum value from each edge value of the edge image detected by the first detection means;
Third detection means for detecting a white line candidate point from the positive maximum value and the negative minimum value obtained by the second detection means;
A fourth detecting means for detecting a white line candidate by detecting a straight line from the white line candidate points detected by the third detecting means;
A first estimation unit that obtains a certainty factor of the white line candidate detected by the fourth detection unit and estimates the white line candidate as a white line when the certainty factor of the white line candidate is equal to or greater than a predetermined value;
The contour correction by the output means is performed from the luminance around the white line candidate detected by the fourth detection means by any of the rising edge enhancement, the falling edge enhancement, and both the rising edge and the falling edge enhancement. Second estimating means for estimating whether or not
Invalidity means for invalidating white line estimation by the first estimation means when contour correction is not estimated by the second estimation means, and the road surface brightness of the image captured by the imaging means tends to be saturated; ,
The white line estimation apparatus characterized by comprising.

Images