JP2003223700A - Method for judging preceding vehicle - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To avoid losing preceding vehicles owing to the fact that a white line can not be detected and to perform the accurate selection of the preceding vehicles using the effect of the detection and judgment of a white line by fusion. <P>SOLUTION: A preceding vehicle on own vehicle lane is selected from forward objects by collating the position of own vehicle 3 on a travel lane obtained by processing image from a camera 2 with the position of the object detected by a scanning laser radar 1. If the reliability of the white line detected by the image processing is low or the white line cannot be detected, the cause of lowered reliability or of the lost white line detection is presumed. A preceding vehicle is selected from the forward objects detected by the scanning laser radar 1 after judging the direction in which own vehicle lane exists according to the cause. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a preceding vehicle judging method for judging a vehicle (preceding vehicle) traveling in front, and in particular, judges a preceding vehicle without being undecidable even during white line detection lost. A method for determining a preceding vehicle.

[0002]

2. Description of the Related Art A conventional invention for detecting a preceding vehicle using a white line detection result obtained by image processing and a vehicle position measurement result obtained by a radar is disclosed in, for example, Japanese Patent Laid-Open No. 2001-93098. In this method, each white line detected by image processing is given a value (reliability value) indicating the likelihood of own lane, and a vehicle existing at a position with a high probability of own lane is determined as a preceding vehicle. It is characterized in that it is possible to cope with a lane change by exchanging the determination of the vehicle on the lane and the vehicle on the adjacent lane according to the lane probability.

[0003]

[Problems to be Solved by the Invention] However, this method is
Although it is a proposal for enabling detection of a preceding vehicle in a scene where it is easy to not detect a white line such as a lane change, it does not correspond to the case where the lane cannot be detected.

Although the reliability is taken into consideration, even if the reliability is low, the white line shape detected with low reliability is used for the selection of the preceding vehicle, so that the detection result is actually incorrect. In this case, even if the preceding vehicle is detected in the distance measurement, the white line integration causes the vehicle existing in the wrong direction to be wrongly selected as the preceding vehicle due to the integration with the white line. There are also adverse effects.

[0005]

In order to achieve the above object, the invention according to claim 1 is mounted on a vehicle, and a step of measuring a distance and a direction to a preceding vehicle in front of the vehicle, and a vehicle-mounted camera. A step of detecting a white line shape in front of the image photographed in, a step of determining the reliability of the result of detecting the white line shape, and a step of determining the cause of the change in reliability when the reliability changes. And a step of obtaining the position of the preceding vehicle on the basis of the cause of the change in reliability when the reliability is lower than a predetermined threshold value.

According to a second aspect of the present invention, in the preceding vehicle determination method according to the first aspect, when it is determined that the cause of the reliability change is a lane change of the own vehicle, it is newly detected after the lane change. If the leading vehicle is determined based on the white line shape, and the cause of the change in reliability is determined to be a sharp curve, the preceding vehicle measured immediately before the change in reliability is determined to be the preceding vehicle. , Is characterized.

According to a third aspect of the present invention, in the preceding vehicle determination method according to the second aspect, as a step of determining that the cause of the reliability change is a lane change, the Y of the own vehicle with respect to a white line is determined.
The case is characterized in that the reliability decreases after the aw angle continues to increase, or the reliability decreases after the distance in the vehicle lateral direction with respect to the white line continues to change.

According to a fourth aspect of the present invention, in the preceding vehicle determination method according to the second aspect, the reliability of the white line is increased after the curvature of the white line continues to increase as a step of determining the cause of the reliability change to be a sharp curve. When it is decreased, it is characterized by.

According to a fifth aspect of the present invention, in the preceding vehicle determination method according to the second aspect, if it is determined that the cause of the reliability change is a lane change of the own vehicle, from the start to the end of the lane change. During the period, both the preceding vehicle before the lane change and the preceding vehicle after the lane change are determined as the preceding vehicles, and a flag indicating that the preceding vehicle before the lane change and the preceding vehicle after the lane change is attached is characterized. And

According to a sixth aspect of the present invention, in the method for determining the preceding vehicle according to the fifth aspect, as a step of determining the end of the lane change, after the Yaw angle of the vehicle with respect to the white line continues to increase. It is characterized in that it is set to a time point when the vehicle starts to decrease, or a time point when the lateral distance from the white line on the lane changing side to the vehicle continues to decrease and then increases.

According to a seventh aspect of the present invention, in the lane change according to the fifth aspect, as a step of determining the end of the lane change, the end point of the white line in the screen edge direction is located substantially in the center on the lower side of the screen. Or if the inclination of the white line is substantially vertical to the screen.

According to an eighth aspect of the present invention, in the preceding vehicle determination method according to the second aspect, when it is determined that the cause of the change in reliability is a sharp curve, the vehicle is judged based on the reliability value. It is characterized in that the expected shape of the white line assumed from the steering angle or yaw rate of the vehicle and the vehicle speed is switched to the white line shape detected from the vehicle-mounted camera, and the preceding vehicle is determined from these results.

According to a ninth aspect of the present invention, in the preceding vehicle determination method according to the second aspect, when the cause of the change in reliability is determined to be a sharp curve, the white line shape detected by the vehicle-mounted camera is detected. When the winker operation in the direction opposite to the curve direction is performed, the cause of the reliability change is changed to lane change.

[0014]

The present invention is intended to solve such a problem, and in claim 1, the position of a radar or the like mounted on a vehicle and capable of measuring the distance and direction to the preceding vehicle in front of the vehicle. Using a measuring device and a camera (imaging device) that can detect the white line shape ahead, based on the white line shape detected by image processing, the position / orientation of the preceding vehicle relative to the lane shape in which the vehicle is running When the relationship is obtained as information including the reliability of the white line shape detection result, the white line shape for determining the own vehicle lane cannot be detected, or the reliability of white line detection falls below a predetermined threshold. Or, when the own vehicle lane cannot be detected, based on the white line shape detected immediately before and the positional relationship between the own vehicle and the preceding vehicle,
The reason why the reliability of the white line shape indicating the own vehicle lane becomes low is determined. Then, based on the cause, the direction in which the lane in which the vehicle is traveling is present is predicted, and the preceding vehicle in that direction is selected as the preceding vehicle in the vehicle lane among the preceding vehicles detected by the radar. The reliability is low,
It is possible to prevent the determination error of the preceding vehicle due to the integration with the white line shape detection result that is highly likely to be erroneously detected. In addition, it is possible to determine the preceding vehicle without being unable to determine even during the white line shape detection lost.

In the second aspect of the present invention, in the selection of the preceding vehicle of the first aspect, when it is determined that the change in reliability of the result of detecting the white line shape is the lane change of the own vehicle, a new detection is made after the lane change. The preceding vehicle is determined based on the determined white line shape, and when the cause of the change in reliability is determined to be a sharp curve, the preceding vehicle measured immediately before that is continuously determined as the preceding vehicle. As a result, for example, on a curved road, the preceding vehicle needs to determine the vehicle on the right (left) side as the preceding vehicle if the vehicle is on the right (left) curve, not the front of the vehicle, but it is detected by the radar. Nevertheless, the preceding vehicle lost due to the white line shape detection lost can be prevented.
Further, when the white line shape is not detected, it is easily conceivable to assume the direction in which the preceding vehicle is present by using the steering angle of the vehicle, but the method using this steering angle is effective when traveling on a curve. Not applicable when changing lanes. For example, in the case of a lane change to the left lane and a left curve, the steering is on the left side, but the direction of presence of the preceding vehicle at that time is on the left side when the vehicle is traveling on the left curve, while on the other hand in the case of a lane change. This is because the lane to the right of the vehicle immediately after the steering is changed to the position of the own lane immediately after the lane change, and the vehicle on the right side is erroneously determined as the preceding vehicle after the lane change. In the method of the present application, the direction in which the preceding vehicle exists is determined according to the cause. Therefore, when the white line is lost due to such two different causes, erroneous determination depending on the steering angle is prevented and the vehicle is present in the correct direction. The vehicle can be selected as the preceding vehicle.

Further, in claim 3, when it is determined that the cause of the reliability change of claim 2 is a lane change, the reliability decreases after the Yaw angle of the own vehicle with respect to the white line continues to increase, or the white line changes. If the reliability decreases after the lateral distance of the vehicle to the vehicle continues to change, it is determined that the cause is a lane change. As a result, the determination of claim 2 can be made more accurately.

Further, in claim 4, when the cause of the change in reliability of claim 2 is determined to be a sharp curve, if the reliability decreases after the curvature of the white line continues to increase, it is caused by a sharp curve running. Judge that there is. Thereby, as in claim 3,
The determination of claim 2 can be made more accurately.

Further, in claim 5, when the preceding vehicle is selected during the lane change of claim 2, both the preceding vehicle before the lane change and the preceding vehicle after the lane change are performed between the lane change start and the lane change end. Is set as the preceding vehicle, and a flag indicating the meaning of the preceding vehicle before the lane change and the meaning of the preceding vehicle after the lane change is attached to each. As a result, it is possible to apply the warning and control using the situation of both lanes. For example, the determination as to whether the vehicle before the lane change or after the lane change should be considered as the preceding vehicle is based on the traffic situation and the driver's sense at that time. If you switch the judgment of the preceding vehicle too early,
If you get too close to the vehicle on the lane you were driving immediately before, or if the timing is too late, there will be problems such as accelerating even if there is a vehicle ahead immediately after changing lanes. Depending on the scene, which one is the target of warning or control while the lane is being changed.
Therefore, by flagging the vehicles on both lanes in which lane and outputting them as the preceding vehicles, it is possible to improve the performance of the control and the alarm device using the detection status of the preceding vehicles. .

Further, in claim 6, when determining the end of the preceding vehicle at the time of changing lanes in claim 5, when the Yaw angle of the own vehicle with respect to the white line continues to increase and then starts to decrease, Alternatively, it is the time when the lateral distance from the white line on the side where the lane is changed to the vehicle continues to decrease and then starts to increase. As a result, it becomes possible to judge the preceding vehicle on the new lane before the lane change is completed, so that it can be applied to control etc. quickly, and acceleration / deceleration at the time of lane change is smooth. You will be able to.

Further, in claim 7, when determining the end of the preceding vehicle at the time of changing the lane in claim 5, when the end point of the white line in the screen edge direction is located at the substantially lower center of the screen, or is detected. It is assumed that the inclination of the white line is substantially vertical to the screen. As a result, the same effect as the sixth aspect can be obtained.

Further, in claim 8, when the reliability of the white line shape detection result is lowered due to the sharp curve of claim 2,
Based on the reliability value, the expected shape of the white line assumed from the steering angle or yaw rate of the host vehicle and the vehicle speed is switched to the white line shape detected by the vehicle-mounted camera, and the preceding vehicle is determined from these results. . As a result, even if the white line cannot be detected due to a sharp curve while traveling on a curve or the reliability is extremely low, if the vehicle ahead can be detected, the preceding vehicle will not be lost. It is possible to correctly track the preceding vehicle (even on a sharp curve where the white line cannot be detected).

In the ninth aspect, the reliability of the white line shape detection result is lowered due to the sharp curve of the second aspect, and the blinker is operated in the direction opposite to the white line shape curve detected from the vehicle-mounted camera. If so, lane changes were the cause of the change in reliability. This makes it possible to correctly determine the lane of the preceding vehicle even when the lane is changed in the direction opposite to the curve while traveling on a curved road.

[0023]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described below with reference to the drawings, taking as an example the case where a scanning laser radar is mounted in a distance measuring device.

<< Description of System Configuration >> FIG. 1 is a system configuration diagram, and FIG. 2 is a mounting position of the scanning laser radar 1 and the camera 2 and the scanning laser radar 1 and the camera 2.
It is explanatory drawing of the reference position system for showing the position of the detection target object with respect to the vehicle carrying. As shown in FIG. 2, all explanations are made by mounting both the optical axis of the vehicle-mounted camera 2 and the scanning center axis of the vehicle-mounted scanning laser radar 1 at the center of the vehicle 3 in a direction and position parallel to the straight traveling direction of the vehicle. It is assumed that the origin of is installed in the same position in the vertical direction on the road surface. However, according to the present invention, even if the optical axis or the central axis is not in the straight traveling direction or the mounting position is deviated from the central axis of the vehicle 3, if the angle and the position are taken into consideration in the geometrical calculation, all of the following are obtained. It consists of an explanation. Further, the position of the detected object with respect to the own vehicle 3 and the position of the white line with respect to the own vehicle 3 are the reference coordinate system XYZ axes (the horizontal position is the X axis, the direction perpendicular to the road surface is the Y axis, and the distance is Z).
Axis).

<< Explanation of Claim 1 >> First, with respect to Claim 1, an example of a mode in which a preceding vehicle is selected will be described along the flow of FIG. In addition, the preceding vehicle in the present specification refers to a vehicle located closest to the front in the lane in which the vehicle is traveling. First,
A vehicle ahead is detected by using the scanning laser radar 1, and the distance to the vehicle and the lateral position (x coordinate) are measured (S101 in FIG. 1). Vehicle detection and its position calculation from the scanning laser radar 1 are described in, for example, Japanese Patent Application Laid-Open No.
A method such as 00-317815 or JP-A-2000-317821 may be applied.

On the other hand, in the image processing, first, the white line is detected (S102 in FIG. 1). In the white line detection, based on the shape of the white line on the image, the Yaw of the host vehicle 3 for the white line is detected.
The angle in the X-axis direction from the straight white line at the corner to the vehicle 3 (hereinafter, referred to as lateral displacement) and the forward curvature of the vehicle lane are obtained.
The Yaw angle, the lateral displacement, and the forward curvature obtained by the white line detection are continuously obtained every time and stored as time-series information (S103 in FIG. 1). For the white line detection for obtaining these values, for example, a method disclosed in Japanese Patent Laid-Open No. 10-011580 may be applied.

Next, the reliability of whether or not the white line indicating the vehicle lane is detected is determined (S104 in FIG. 1).
The reliability is determined, for example, from the number of edges on the white line detected from the image or the shape-likeness of the vehicle lane determined from the shape of the white line. Here, the reliability determination method will be supplementarily described. FIG. 3A is a diagram showing an image when a front white line is picked up by the camera 2 mounted on the vehicle 3 and a state in which the white line on the image is detected and the area of the vehicle lane is detected. is there. Normally, the white line detects an edge on the image, and determines whether the shape of the edge is close to the shape of the front lane. Therefore, the reliability of the white line detection result indicating the own vehicle lane is determined from the number of detected edges and the “white line shape likeness” of the white lines drawn by aligning the detected edges. Edge detection is obtained by various classical image processing methods such as image differentiation.

Usually, in bad weather or when the white line itself is faint, it is difficult to detect edges, and reliability is lowered.
In addition, “white line shape likeness” means that the white lines on both sides of the vehicle lane imaged in front of the vehicle are drawn on the road, and the distance between the left and right sides of the vehicle 3 is about 3.5 m. Since it has characteristics, it is determined whether or not the shape matches the shape. In other words, even in an environment where edges are easily detected, if there are lines other than white lines on the road surface, or if you can see many white lines while changing lanes, etc., the white lines will be outside the angle of view. If the curve is so sharp that it will come out, it may be difficult to determine which edge indicates the shape of the vehicle lane, or there may be problems such as erroneous edges other than the white line on the vehicle lane being mistaken for white lines (Fig. Three
(B)).

In order to solve such a problem, normally,
In the recognition of the own vehicle lane by image processing, the edge detection range is limited to only the range captured by the left and right lanes, or the inclination of the edge detected therein is limited to that indicating the lane, for example, The edge detected on the right side of the image is the lane on the right side of the own vehicle, so that the inclination of the edge is limited to the downward-sloping edge so that only the edge on the white line indicating the own vehicle lane is detected. The device for it is taken (Fig. 4). For example, FIG. 5A shows the shape of the own vehicle lane on the imaged image of the own vehicle lane traveling on a straight road, FIG. 5B shows the shape of the own vehicle lane traveling on a gently curved road, and FIG. (A), (b) is a figure showing the shape of the white line existing ahead while the lane is being changed, and FIG. 6 (c) is a figure showing the shape of the vehicle lane while traveling on a sharply curved road. Among these,
As can be seen from the figure, only the lanes with high vehicle lane reliability can be accurately represented in FIGS. 5A and 5B. In FIGS. 6A, 6B, and 6C, the edge is out of the vehicle lane detection range or is detected as a shape different from the shape of the vehicle lane. In some cases, it is difficult to make a judgment. Therefore, the reliability is determined from the number of detected edge points and the shape of the lane formed by the detected edges, and when the edges cannot be detected due to bad weather, or even in good weather. In an environment where it is difficult to correctly detect the shape of the left and right white lines indicating the vehicle lane, such as during a lane change or a sharp curve, the reliability of detection of the vehicle lane becomes low or the vehicle is lost. The reliability determination in S104 of FIG. 1 is calculated / determined using such a principle.

Next, when the reliability of white line detection is low or when the lane of the host vehicle is determined to be lost, the reliability becomes low based on the situation when the lane was previously determined. The cause is determined (S105). FIG. 5 and FIG. 6 show the detected positions of edges and the positions of detected lanes on the image in various scenes. For example, the number of detected edges and the detected positions on the image and the reliability are low. The previously detected Yaw angle, lateral displacement, curvature, etc. have different characteristics depending on the respective situations. For example, bad weather when there are few edges continuously in time, and gradual lane changes when the number of edges detected is high and reliability is gradually decreasing.In this case, low It is possible to study the driving scene and the cause of low reliability, such as running on a curve if the curvature before is high and running on a straight road if it is low.

Next, the preceding vehicle is selected on the basis of the position of the vehicle ahead detected by the scanning laser radar 1 and the position of the detected lane. When selecting the preceding vehicle, based on the reliability of the white line detection, if the detection reliability of the own vehicle lane is high, the positional relationship between the result detected by the image processing and the object detected by the scanning laser radar 1 is used. If the vehicle in the vehicle lane is the preceding vehicle and the reliability is low, the vehicle existing in the direction determined to be the direction of the vehicle lane is determined as the preceding vehicle depending on the reason why the vehicle lane is no longer detected. Yes (S106). Finally, the position and distance of the vehicle determined to be the preceding vehicle, the reliability of being the preceding vehicle, and the like are output as a result (S107).

<< Explanation of Claim 2 >> Next, an embodiment of claim 2 will be described.

In the selection of the preceding vehicle at the final stage of claim 1, it is determined that the cause of the reduction in the reliability of detection of the own vehicle lane is a lane change as shown in FIGS. 6 (a) and 6 (b). In this case, the direction of the vehicle lane that should be newly detected after the lane change, that is, the direction opposite to the steering angle direction,
When it is determined to be a sharp curve as shown in FIG. 6C, the vehicle existing in the direction of extension of the curved road detected immediately before that, that is, the direction of the steering angle is selected as the preceding vehicle. To do so. FIGS. 7A and 7B are views showing the own vehicle 3 and the steering directions of the own vehicle 3 at the time of changing lanes and while traveling on a curved road, and the positional relationship between the own vehicle 3 and the preceding vehicle at that time. Is. As you can see from the figure,
When the lane is changed to the left lane, steering is directed to the left, while the preceding vehicle immediately after the lane change is on the right side of the center line of the host vehicle. However, in the case of a sharp curve, if it is a left curve,
The steering direction is left, and the position of the preceding vehicle is also left with respect to the center line of the host vehicle. As described above, even when the vehicle lane is lost, the correct preceding vehicle can be determined if the vehicle ahead is detected.

<< Explanation of Claim 3 >> Next, an embodiment of claim 3 will be described.

FIG. 8 shows the change in the Yaw angle of the front white line with respect to the own vehicle when changing lanes, and FIG. 9 shows the change in lateral displacement up to the white line after changing lanes. The route when changing lanes depends on the driver and road conditions, but as shown in Fig. 8, Y
After the aw angle gradually increases, the size of the yaw angle starts to decrease after the lane change. Alternatively, as shown in FIG. 9, the lateral displacement gradually decreases, and at the intermediate point, the white line is outside the detection range as shown in FIG. 6B and lateral displacement = A4 image, and when the lane change is completed, A white line newly appears in the detection range, and the lateral displacement with the white line gradually increases.
There are characteristic changes in the aw angle and lateral displacement. The tendency of the change of this value, that is, after the Yaw angle gradually increases,
Alternatively, if the detection reliability of the vehicle lane becomes low after the lateral displacement gradually becomes shorter, it can be determined that the cause is a lane change.

As can be seen from FIGS. 8 and 9, the change in the Yaw angle and the lateral displacement of the host vehicle 3 with respect to the white line is not limited to the Yaw angle and the lateral displacement calculated from the white line detection result, but also on the image. It can also be obtained from the change in the lane itself detected in. For example, from the state of the images shown on the right side of each of FIGS. 8 and 9, for example, only the white line on the right side detected on the image is continuously focused, and the inclination of the white line gradually becomes a vertical angle. After the lane change, it is possible to determine that there is a lane change because the vehicle is moving to the left. Similarly, paying attention to the lane on the right side and observing the movement of the lowest position (the position shown by a star in FIG. 9) on the image where the lane is detected, not the inclination, the lane change can be made. You can see what happened.

<< Explanation of Claim 4 >> Next, an example of the embodiment of Claim 4 will be described.

When entering a sharp curve from a straight road, the front curvature gradually increases. Usually, as shown in FIGS. 5 and 6, when the curve is too steep and goes out of the angle of view, it is difficult to accurately determine the own vehicle lane. Processing such as stopping lane detection and switching to lost determination is performed. However, this loss does not occur suddenly and the curvature gradually increases, so it is possible to capture a numerical change in which the curvature gradually increases until the curvature exceeds the processing value. That is, by observing the value of the curvature obtained by the detection of the white line and the state of its change, it is possible to determine that the loss is due to the sharp curve.

<< Explanation of Claim 5 >> Next, an example of the embodiment of Claim 5 will be described.

When selecting the preceding vehicle when the reliability of the white lane is reduced due to the lane change, which of the lanes before and after the lane change should be judged as the own lane is determined from the start of the lane change to the end of the lane change. It varies depending on the road conditions of the place, the driver's feeling, and the method of use for control and display / warning. In order to respond to these changes, when it is determined that the lane has changed, both vehicles existing in the lanes before and after the lane change are considered to be the preceding vehicles, and the vehicles on the lanes before and after the lane change are assigned to the respective vehicles. Output with a flag indicating that there is. Of the two lanes detected during the lane change, which is the lane before the lane change and which is the lane after the lane change is determined by the Yaw shown in FIGS. 8 and 9.
From the direction of increase or decrease in the corner or lateral displacement, it is possible to determine whether the lane change is to the right or the left. For example, in FIGS. 8 and 9, the displacement to the right lane continues to decrease, so it is possible to determine the lane change to the right.

<< Explanation of Claim 6 >> Next, an example of the embodiment of Claim 6 will be described.

According to the second aspect of the present invention, the timing for determining the lane change while the lane is changing is the time when the Yaw angle of the own vehicle with respect to the white line which continues to increase or the time when the Yaw angle of the own vehicle continues to decrease. It is assumed that the lateral displacement of the vehicle with respect to the white line begins to increase. As shown in FIGS. 8 and 9, the change point of the increase or decrease of the Yaw angle or the lateral displacement corresponds to the change point of the lane. For this reason, the timing of lane switching can be accurately and promptly determined by taking into consideration the increase and decrease of the Yaw angle and the lateral displacement.

<< Explanation of Claim 7 >> Next, an example of the embodiment of Claim 7 will be described.

The instant of lane switching can be determined not only from changes in the Yaw angle and lateral displacement, but also from the position of the edge detected on the image. For example, as described above, when changing the right lane, the position where the bottom edge (star in the figure) of the right white line on the image is detected is from the right side in the image to the left side via the bottom edge of the image. You can see the phenomenon of moving to. It can also be seen as a phenomenon in which the inclination of the white line changes from a right downward direction to a vertical direction and then changes to a left downward edge. That is, the lane change can be determined by simultaneously detecting the shape of the traveling lane of the own vehicle, focusing on the movement of a certain white line, and confirming the presence or absence of such a phenomenon by image processing. In this case, from the result of this image processing, the lane switching is performed at the time when the bottom edge position of a certain lane is detected at the center of the bottom edge of the image, or the inclination of the edge detected as the lane becomes almost vertical. It can be determined that it is time.

<< Explanation of Claim 8 >> Next, an example of the form of Claim 8 will be described.

When calculating the curve road extension direction performed when the reliability of white line detection is lowered due to a sharp curve in claim 2, it is assumed from the steering angle or yaw rate of the host vehicle 3 and the vehicle speed. The curve drawn by the trajectory of the vehicle 3 is defined as the vehicle lane. The own vehicle lane can be calculated from the steering angle or yaw rate and the vehicle speed by applying the own vehicle trajectory as the lane shape. The trajectory of the own vehicle can be obtained by considering the steering angle, the speed, and the slip angle generated at that time from the equation of motion of the vehicle. In addition, on a curved road, the white line detected from the image is used for the front lane shape or the lane shape estimated from the trajectory obtained from the steering angle or yaw rate and the vehicle speed is used to determine the reliability of the white line detection. Depending on the value, for example, when the reliability is equal to or higher than a predetermined threshold value, the image processing result is obtained, and when the reliability is equal to or lower than the predetermined threshold value, an estimated lane shape obtained from the steering angle and the speed may be used.
As described above, the reliability calculation method may be empirically obtained from the number of detected edges and the detected curvature of the front lane (such as low in the case of a sharp curve).

<< Explanation of Claim 9 >> Finally, an example of the embodiment of Claim 9 will be described.

In selecting a traveling vehicle while traveling on a sharp curve in claim 8, when a blinker appears in a direction opposite to the curve obtained from the image or the steering angle and the vehicle speed, it is determined that the lane is changed during traveling on a curved road. The determination at the time of changing lanes according to claims 5 to 7 is applied. Whether the curve is facing left or right can be determined from the shape of the lane detected by image processing. Further, it can be determined from the direction in which the turn signal is coming out or the signal line in the vehicle coming out of the turn signal. By comparing these values, the direction of the curve and the direction of the lane change can be determined.

By taking this into consideration, it becomes possible to cope with a lane change during a curve, and in the case of a lane change caused by determining that the direction of the curve is the existing position of the preceding vehicle in the curve road proposed by the present invention. It is possible to eliminate the adverse effect of the above, and it is possible to make use of the judgment when changing lanes.

<< Effects of Each Claim >> As described above, according to claim 1, a position measuring device such as a scanning laser radar 1 mounted on a vehicle and capable of measuring the distance and direction to an object existing in front of the vehicle. And the position of the traveling lane of the vehicle 3 detected by image processing and the position of the object detected by the scanning laser radar 1 are collated by using an imaging device such as a camera 2 capable of detecting the front white line. In a device that correctly selects a preceding vehicle on the vehicle lane from among objects detected by the scanning laser radar 1, the reliability of white line detection detected by image processing is low,
Alternatively, when the white line is not detected, it is determined whether the reliability is low or the white line detection is lost, and in accordance with the cause, the direction in which the vehicle lane exists is determined.
The preceding vehicle is selected from the objects in front detected by the scanning laser radar 1. As a result, the problem that occurred in the fusion process (Fusion Process) between the image processing and the scanning laser radar 1, for example, the scanning laser radar 1 could not detect the white line even though the preceding vehicle was detected. It becomes possible to avoid the adverse effect of the preceding vehicle lost due to the above, and it becomes possible to more accurately select the preceding vehicle by utilizing the effect of the white line detection determination by the fusion.

In the second aspect of the present invention, in the selection of the preceding vehicle according to the first aspect, when it is determined that the lane change causes a decrease in the reliability of white line detection, the direction of the lane to be newly detected after the lane change is completed, When it is determined that the vehicle is a sharp curve, the vehicle existing in the direction of the curve road extension detected up to that point is selected as the preceding vehicle. The direction in which the preceding vehicle exists is the steering direction when driving on a curve,
When the lane is changed, it exists in the opposite direction to the steering and in the opposite direction to the movement of the host vehicle 3. Therefore, when there is no white line information ahead, when selecting a preceding vehicle by a commonly used method, for example, a method using the steering angle applied to the determination of the preceding vehicle, the actual preceding vehicle exists when the lane is changed. The vehicle in the direction opposite to the direction in which the vehicle will be driven is determined to be the preceding vehicle. In this method, the direction in which the preceding vehicle exists is determined according to the cause. Therefore, by considering the white line lost due to such two different causes, erroneous determination depending on the steering angle is prevented, and the vehicle is in the correct direction. It becomes possible to select the vehicle to be operated as the preceding vehicle.

According to the third aspect, when the cause of the decrease in reliability of the white line detection of the first aspect is determined, the Yaw for the white line is determined.
If the reliability decreases after the corner increases or the lateral displacement of the white line with respect to the host vehicle 3 continues to decrease, it is determined that the cause is a lane change. As a result, the determination of claim 2 can be made more accurately. In particular, in this method, the lane change is judged using the image processing device already installed. There is an advantage that a device only for lane change determination is unnecessary. Further, the lane change determination using the Yaw angle and the lateral displacement obtained from the white line detection result is performed by the white line detection processing only from the movement on the edge detected at the position properly determined to be the lane. Since the lateral displacement will be calculated, for example,
Comparing the movement relationship with the lane more reliably than methods such as detecting the moving direction of the vehicle from the optical flow that uses all edges in the image, including edges not necessarily on the lane on one image There is also the effect that you can. In addition, in the case of lane area detection, the area where the white line is detected may be limited, so the lane at the originally visible position may not be detected, but it is detected with a highly reliable value for lane change determination. Always keep an eye on the movement of the displayed lane and keep track of that lane even if it leaves the detection area. The lane change can also be determined by observing the change in the position. Even in the case of this method, only the movement of the edge existing on the road surface that is confirmed to be a white line is tracked. Therefore, as in the case of using the Yaw angle or the lateral displacement obtained as a result of the white line detection, the lane is accurately measured. You can judge changes. This method also has an advantage that the lane change can be stably determined even when it becomes difficult to detect the lane that is not being tracked.

Further, in claim 4, when the cause of the decrease in reliability of white line detection in claim 1 is determined, if the reliability of white line detection becomes low after the curvature of the white line continues to increase, a sharp curve is generated. The configuration is such that it is determined that the vehicle is running. As a result, similarly to the third aspect, it is possible to perform the curve determination with the existing configuration without a special device for determining the curve traveling. Further, as in the third aspect, this determination is also based on the movement of the edge determined to be the own vehicle lane, so that the road surface determination processing such as curve detection or texture analysis using an image at a certain point of time, etc. There is also an effect that the change in the shape of the lane can be obtained more reliably than the calculation of the curvature based on.

Further, in claim 5, when selecting a preceding vehicle when changing lanes according to claim 2, between the start of the lane change and the end of the lane change, it is detected in the lane before the lane change and in the direction after the lane change. Both of the vehicles existing at the proper positions are set as preceding vehicles, and a flag indicating the meaning before the lane change and the meaning after the lane change is attached to each of them. As a result, it is possible to apply the warning and control using the situation of both lanes. For example, if you switch the judgment of the preceding vehicle too early, you will get too close to the vehicle on the lane that was running immediately before, or if the switching timing is too late, the vehicle will accelerate even though there is a vehicle ahead immediately after the lane change. Although there are problems such as being lost, because the situation of both lanes is known,
It can also be linked to control using the detection status of the preceding vehicle and performance improvement of the alarm device.

Further, in claim 6, when the preceding vehicle is selected at the time of changing lanes according to claim 2, the judgment of the own vehicle lane change keeps increasing, and the Yaw angle of the own vehicle with respect to the white line starts to decrease. The time point was set to be the time point, or the time point when the lateral displacement of the own vehicle, which continued to decrease, started to increase. This makes it possible to determine the preceding vehicle on the new lane at the moment of lane change, that is, before the lane change is completed, so control such as speed depending on the distance to the preceding vehicle at the time of lane change can be performed quickly. This also has the effect of facilitating acceleration / deceleration when changing lanes.

Further, in claim 7, when the preceding vehicle is selected while the lane is being changed in claim 2, the judgment of the lane change of the own vehicle is continued to move toward the center of the image, and a white line is detected in the image. The configuration is such that the time is closer to the center than the predetermined position, or the time when the inclination of a certain white line is almost vertical. As a result, similarly to the sixth aspect, it becomes possible to quickly determine the preceding vehicle at the time of changing lanes, which facilitates application to control.

Further, in claim 8, when the preceding vehicle is selected while traveling on a sharp curve according to claim 2, the curve drawn by the trajectory of the host vehicle 3 assumed from the steering angle of the host vehicle 3 and the vehicle speed is drawn in the host vehicle lane. As a configuration, depending on the reliability value of the white line detection, the position of the vehicle lane is switched between the case where the white line detection result is obtained and the case where the position is obtained from the curvature obtained from the steering angle or yaw rate and the vehicle speed. As a result, even if the white line cannot be detected due to a sharp curve or the reliability is extremely low during traveling on a curve, if the vehicle ahead can be detected, the traveling direction of the host vehicle 3 is changed. Since the vehicle in front of the vehicle can be determined to be the preceding vehicle, it is possible to prevent the preceding vehicle lost due to the white line detection lost.

Further, in claim 9, when the blinker is instructed in the direction opposite to the curve obtained from the image or the steering angle and the vehicle speed in selecting the preceding vehicle during traveling on a sharp curve in claim 8, It is configured to apply the determination at the time of changing lanes of items 3 to 6. According to the present embodiment, even if the vehicle lane is not detected by determining whether the vehicle is traveling on a curved road or changing the lane, it is possible to correctly select the preceding vehicle without adversely affecting the white line detection lost. Further, in the ninth aspect, it is possible to prevent an adverse effect due to the feature of the present invention by considering the lane change during the curve running.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a system to which a preceding vehicle determination method according to the present invention is applied.

FIG. 2 is an explanatory diagram of mounting positions of a scanning laser radar and a camera on a vehicle and setting of a coordinate system.

FIG. 3 is an explanatory diagram of an image pickup position of a front white line and an own vehicle lane area on an image.

FIG. 4 is an explanatory diagram of white line detection areas on both sides of the vehicle lane on the image.

FIG. 5 is an explanatory diagram showing how white lines are detected in various road shapes and vehicle postures.

FIG. 6 is an explanatory diagram of a state of a white line detection result in various road shapes and vehicle postures.

FIG. 7 is a diagram showing a direction in which a preceding vehicle exists with respect to the own vehicle at the time of changing lanes and traveling on a curve.

FIG. 8: Y of the own vehicle with respect to the white line at the time of sudden lane change
It is a figure showing the change of aw angle or lateral displacement.

FIG. 9 is a diagram showing a change in Yaw angle or lateral displacement of the host vehicle with respect to a white line when a lane is gently changed.

[Explanation of symbols]

1 scanning laser radar 2 camera 3 vehicles

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 2F065 AA01 AA31 AA54 BB02 BB05                       BB15 CC12 CC40 FF04 FF11                       GG04 JJ03 JJ05 JJ26 LL61                       MM16 QQ28 QQ31 TT08                 3D044 AA25 AB01 AC31 AC56 AC59                       AE01 AE04 AE15                 5H180 AA01 CC03 CC04 LL04                 5J084 AA05 AA10 AB01 AC02 BA03                       BA48 DA01

Claims (9)

[Claims] 1. A step of measuring a distance and an azimuth of a preceding vehicle which is mounted on a vehicle and located ahead of the vehicle, a step of detecting a white line shape in front of an image captured by an in-vehicle camera, and a step of detecting the white line shape. A step of determining the reliability of the result, a step of determining the cause of the change in reliability when the reliability changes, and a step of determining the reliability of the reliability when the reliability is lower than a predetermined threshold value. A method of determining a preceding vehicle, comprising the step of obtaining the position of the preceding vehicle based on the cause of the change. 2. The preceding vehicle determination method according to claim 1, wherein when it is determined that the cause of the reliability change is a lane change of the own vehicle, the preceding line is detected based on a white line shape newly detected after the lane change. When a vehicle is determined and it is determined that the cause of the change in reliability is a sharp curve, the preceding vehicle measured immediately before the change in reliability is continuously determined to be a preceding vehicle. Car judgment method. 3. The method of determining a preceding vehicle according to claim 2, wherein the step of determining that the cause of the reliability change is a lane change, the reliability after the Yaw angle of the host vehicle with respect to the white line continues to increase. The following is a method of determining a preceding vehicle, characterized in that the reliability decreases after the distance in the vehicle lateral direction continues to change with respect to the white line. 4. The preceding vehicle determination method according to claim 2, wherein the step of determining the cause of the change in reliability as a sharp curve is a case where the reliability decreases after the curvature of the white line continues to increase. A preceding vehicle determination method characterized by the above. 5. The preceding vehicle determination method according to claim 2, wherein when it is determined that the cause of the reliability change is a lane change of the own vehicle, the lane change is performed from the start to the end of the lane change. Both the preceding vehicle and the preceding vehicle after the lane change are determined as the preceding vehicles, and a preceding vehicle determination method is characterized in that a flag indicating a preceding vehicle before the lane change and a preceding vehicle after the lane change is added. 6. The method of determining a preceding vehicle according to claim 5, wherein, as a step of determining the end of the lane change, the Yaw angle of the vehicle with respect to the white line continues to increase and then starts to decrease. Alternatively, the preceding vehicle determination method is characterized in that after the lateral distance from the white line on the side where the lane is changed to the vehicle continues to decrease, the time is increased. 7. The lane change according to claim 5, wherein, as a step of determining the end of the lane change, when an end point of the white line in the screen edge direction is located in the approximate center on the lower side of the screen, or A preceding vehicle determination method, characterized in that the inclination is substantially vertical to the screen. 8. The preceding vehicle determination method according to claim 2, wherein when the cause of the change in reliability is determined to be a sharp curve, the steering angle or yaw rate of the host vehicle and the vehicle speed are determined based on the reliability value. While switching the expected shape of the white line assumed from, and the white line shape detected from the vehicle-mounted camera,
A preceding vehicle determination method characterized by determining a preceding vehicle from these results. 9. The leading vehicle determination method according to claim 2, wherein when the cause of the change in reliability is determined to be a sharp curve, the winker is in a direction opposite to a white line-shaped curve detected by the vehicle-mounted camera. A method of determining a preceding vehicle, characterized in that, when an operation is performed, the cause of the change in reliability is changed to lane change.