JP2014219960A - Track detection device and track detection method - Google Patents

Abstract

A travel path detection apparatus capable of stabilizing a travel path estimation result is provided. A runway detection device includes an edge detection unit that detects edge points L11, L12, L13, L14, R11, R12, R13, and R14 that are road luminance change points, and a left side that is detected by the edge detection unit. The midpoint calculation unit that calculates the midpoints of the edge points L11, L12, L13, and L14 of the right side and the right edge points R11, R12, R13, and R14 detected by the edge detection unit as reference points that are candidates for the travel path A midpoint position voting unit for voting the position of the midpoint calculated by the midpoint calculation unit, and a position where the number of votes is high among the positions of the midpoints voted by the midpoint position voting unit And a travel path detection unit that detects as [Selection] Figure 5

Description

  The present invention relates to a travel path detection device and a travel path detection method for detecting a travel path of a vehicle.

  Japanese Patent Laying-Open No. 2006-268199 (Patent Document 1) discloses an image processing system that detects a lane mark on a traveling path of a vehicle. This image processing system sets a reference region corresponding to a road surface portion other than a lane mark in a road surface image, creates a luminance histogram by measuring the frequency of luminance in the reference region, and creates a width and a height from the created histogram. A group of portions whose height or area is equal to or greater than a threshold is extracted as a road surface cluster. Further, all white line candidate edge points are detected as primary white line candidate edge points, and those that overlap the reference area among the primary white line candidate edge points are detected as secondary white line candidate edge points.

  Of the secondary white line candidate edge points, only those that do not include the brightness parameter value in the brightness range of the road surface cluster are detected as true white line edge points. In this way, among the secondary white line candidate edge points, those included in the brightness range of the road surface cluster are excluded from the true white line candidates, so that only the true lane mark is accurately detected and erroneous detection of the lane mark is performed. It is suppressed.

JP 2006-268199 A Japanese Patent No. 3684395 Japanese Patent Laid-Open No. 07-128059

  However, the image processing system as described above detects white line candidates as lines as a set of white line candidate edge points. That is, since the lane mark is detected as a line on the image, when the actual lane mark is a composite line or has a branch line, it may not be possible to determine which line is a desired white line on the image. is there. Therefore, there is a problem that the estimation result of the lane becomes unstable. In addition, when the actual lane mark is hidden by another vehicle or the like, the white line candidate may not be detected, and in this case as well, the lane estimation result may become unstable.

  Then, the subject of this invention is providing the runway detection apparatus and runway detection method which can stabilize the estimation result of a runway.

  That is, the running road detection device of the present invention includes an edge detection unit that detects an edge point that is a change point of the brightness of the road, and a vehicle running from the left edge point and the right edge point on the road detected by the edge detection unit. A reference point calculation unit that calculates a reference point that is a candidate for a road, a reference point position voting unit that votes the position of the reference point calculated by the reference point calculation unit, and a reference point voted by the reference point position voting unit A travel path detection unit that detects the travel path of the vehicle from the position.

  According to this invention, the reference point calculation unit calculates the reference point from the left edge point and the right edge point, and the reference point position voting unit votes the reference point. A position with a large number is detected as a travel path by the travel path detection unit. In this way, in the present invention, the travel path is detected using the positions of the reference points obtained from the left and right edge points without detecting the white line candidate as a line as in the prior art. Therefore, even when the actual lane mark is a composite line, the travel path is detected by the voting results of the reference points obtained from the edge points detected from the composite lines on both sides, so the travel path detection result is stable. Can be made.

  In the runway detection apparatus according to the present invention, the reference point is preferably a midpoint between the left edge point and the right edge point. In this case, since the reference point calculation unit calculates the midpoint between the left edge point and the right edge point, it is possible to detect a travel route close to the lane center.

  In the runway detection device according to the present invention, the edge detection unit includes a rising edge point that increases in brightness when the detection position is moved from one side to the other side on the captured image of the road, and left and right on the captured image. And detecting a falling edge point at which the brightness decreases when the detection position is moved from one side to the other side of the first reference point, and the reference point calculation unit calculates the reference point from the left rising edge point and the right falling edge point. The reference point is preferably calculated from the left falling edge point and the right rising edge point. In this case, since the reference points for the rising edge point and the falling edge point are calculated as needed, the number of votes for the desired reference point position can be increased. Therefore, since the estimated reference point position can be brought closer to the desired travel path of the lane, the travel path estimation result can be further stabilized.

  In addition, when a distance calculation unit that calculates the distance between the left edge point and the right edge point, and the difference between the average or median distance calculated by the distance calculation unit and the lane width of the road is equal to or greater than a threshold value A pitching angle change calculation unit that calculates a change amount of the pitching angle of the vehicle using the average value or the median value and the road lane width, and the distance calculation unit includes the average value or the median value and the road lane width. Is equal to or greater than the threshold value, the reference point calculation unit corrects the pitching angle of the vehicle by the amount of change in the pitching angle calculated by the pitching angle change calculation unit before calculating the reference point, and again. It is preferable to calculate the distance between the edge points.

  Actually, pitching occurs due to sudden braking while the vehicle is running. However, when pitching occurs in this way, the distance between the edge points on the captured image and the actual width of the traveling path differ. End up. That is, when pitching occurs, the travel path on the captured image is distorted, so that the position of the reference point cannot be calculated correctly, and the estimation result of the travel path may become unstable. However, in the present invention, when the pitching occurs, the pitching angle is corrected and the distance between the edge points is calculated again. Therefore, the influence of the captured image being distorted by the pitching is eliminated, and even when the pitching occurs, the travel path The estimation result can be stabilized. In the runway detection apparatus according to the present invention, the edge point may be an edge point of a white line on the road.

  The runway detection method of the present invention includes an edge detection step for detecting an edge point that is a change point of the brightness of a road in a captured image obtained by imaging a road on which a vehicle is traveling, and a left edge point on the road detected in the edge detection step. A reference point calculating step for calculating a reference point that is a candidate for a vehicle travel path from the edge point on the right side, a reference point position voting step for voting the position of the reference point calculated in the reference point calculating step, and a reference point position And a travel path detection step of detecting the travel path of the vehicle from the position of the reference point voted in the voting process.

  According to the present invention, the reference point is calculated from the left edge point and the right edge point in the reference point calculation step, and the reference point is voted in the reference point position voting step. Many positions are detected as travel paths in the travel path detection step. In this way, in the present invention, the travel path is detected using the positions of the reference points obtained from the left and right edge points without detecting the white line candidate as a line as in the prior art. Therefore, even when the actual lane mark is a composite line, the travel path is detected based on the voting results of the reference points obtained from the edge points detected from the composite lines on both sides, so the detection result of the travel path is stabilized. be able to.

  According to the present invention, it is possible to stabilize the estimation result of the traveling road.

It is a block diagram showing a runway detection device concerning a 1st embodiment of the present invention. It is a perspective view which shows the rising edge and falling edge on the white line of an actual travel path. It is a figure explaining the vote of a midpoint position. It is a perspective view which shows the rising edge and falling edge in a compound line. It is a figure explaining the vote of the midpoint position in a compound line. It is a flowchart which shows the process which estimates the travel path of the travel path detection apparatus of FIG. It is a figure explaining the distance between the edge points on a picked-up image. It is a figure explaining the vote of the midpoint position in the case of a branch line or when a white line is hidden. It is a block diagram which shows the track detection apparatus which concerns on 2nd Embodiment of this invention. It is a histogram which shows the frequency of the value of the half of the vehicle width on a picked-up image. It is a flowchart which shows the process which estimates the travel path of the travel path detection apparatus which concerns on 2nd Embodiment of this invention.

  Embodiments of the present invention will be described below with reference to the drawings. In the following description, the same or equivalent elements will be denoted by the same reference numerals, and redundant description will be omitted.

(First embodiment)
As shown in FIG. 1, the travel path detection apparatus 1 according to the first embodiment is an apparatus that detects a travel path while the vehicle is traveling. The road detection device 1 is used, for example, in a lane keeping assist (LKA) that alerts the driver or supports a steering operation when it is likely to deviate from the road to prevent a lane departure.

  The travel path detection apparatus 1 includes an edge detection unit 21 that detects an edge point that is a change point of the brightness of a road, and a midpoint that calculates a midpoint between a left edge point and a right edge point as a reference point for traveling. The calculation unit (reference point calculation unit) 22, the midpoint position voting unit (reference point position voting unit) 23 for voting the position of the midpoint, and a position with a large number of votes among the midpoint positions is used as the vehicle travel path. And a travel path detection unit 24 for detection.

  The travel path detection device 1 also includes a camera 10 that captures the road being traveled, an ECU (Electronic Control Unit) 20 that performs image processing of an image captured by the camera 10 and detects the travel path, and an ECU 20. And an output unit 30 that receives a control signal and assists a warning for a driver of the vehicle and a steering operation of the vehicle.

  The camera 10 has a function of photographing the front of the road on which the vehicle is traveling, and the photographed image is output to the ECU 20. The ECU 20 performs image processing of a captured image captured by the camera 10, and specifically converts the captured image captured by the camera 10 into a planar view image to detect a traveling path. The edge detection unit 21, the midpoint calculation unit 22, the midpoint position voting unit 23, and the travel path detection unit 24 described above are provided in the ECU 20, for example.

  The edge detection unit 21 detects an edge point that is a luminance change point from the captured image of the road captured by the camera 10. As shown in FIG. 2, when the white line H1 and the white line H2 are on both sides in front of the vehicle, the edge detection unit 21 detects the rising edge points L1 and R1 and the falling edge points L2 and R2 from the captured image. To detect. Here, the rising edge points L1 and R1 are edge points that increase in luminance when the detection position is moved from one side to the other side on the captured image, and the falling edge points L2 and R2 are the other side from the one side on the captured image. This is an edge point where the brightness decreases when the detection position is moved to the side.

  The midpoint calculator 22 calculates the midpoint position between the left edge point and the right edge point at the same horizontal position of the captured image detected by the edge detector 21. Specifically, as shown in FIG. 3, the midpoint calculation unit 22 calculates the middle of the left rising edge point L1 in the captured image converted into the planar view image and the right falling edge point R2 in the captured image. The midpoint between the point and the left falling edge point L2 in the captured image and the right rising edge point R1 in the captured image is calculated. Thus, by calculating the midpoint between the left rising edge point L1 and the right falling edge point R2 and the midpoint between the left falling edge point L2 and the right rising edge point R1, the lane center is further calculated. It is possible to detect a road close to.

  The midpoint position voting unit 23 generates runway candidate sections C1 and C2 as shown in FIG. 3 from the midpoint position calculated by the midpoint calculation unit 22, for example. The runway candidate parts C1 and C2 are displayed in such a manner that the midpoint position calculated by the midpoint calculation part 22 is darkest and gradually becomes thinner toward the periphery, and the runway candidate parts C1 and C2 are displayed in a blurred manner in this way. This avoids a situation where the travel route is extremely shifted. The midpoint position voting unit 23 repeatedly generates the runway candidate units C1 and C2, and repeats the voting of the midpoint position calculated by the midpoint calculation unit 22, for example, the frequency of the midpoint position as shown in FIG. Is generated. Here, the vote of the midpoint position means that the midpoint position calculated by the midpoint calculation unit 22 is recorded as one candidate for the travel path.

  The travel path detection unit 24 detects a position having a large number of votes among the midpoints voted by the midpoint position voting unit 23 as a travel path of the vehicle. The travel path detection unit 24 also detects a position with a high frequency from the histogram generated by the midpoint position voting unit 23 as travel path information, and outputs the detected travel path information to the output unit 30.

  The output unit 30 receives the information on the travel path output by the travel path detection unit 24 and assists the driver in driving the vehicle. Specifically, the output unit 30 may include, for example, an alarm device that warns the driver by voice, image display, or the like when the vehicle deviates from the traveling road. Steering control for assisting the steering of the vehicle and driving the vehicle to the correct traveling path may be performed when the vehicle deviates from the traveling path.

  As described above, in the traveling path detection device 1, as shown in FIG. 3, the midpoint calculation unit 22 calculates the midpoint between the left edge points L 1 and L 2 and the right edge points R 1 and R 2 of the captured image of the camera 10. The point position voting unit 23 votes the middle point, and among the voted midpoint positions, a position with a large number of votes is detected by the traveling path detection unit 24 as a traveling path. In the present embodiment, the white line candidates are not detected as lines as in the prior art, but the midpoint positions of the edge points on the left and right sides are detected as the travel path. Therefore, even when the actual lane mark is a composite line, the travel path is detected by the midpoint voting result between the edge points detected from the composite lines on both sides, so the detection result of the travel path is stabilized. Can do.

  Specifically, as shown in FIG. 4, in the case of a composite line in which two white lines H11 and H12 extend forward on the left side and two white lines H13 and H14 extend forward on the right side. The edge detection unit 21 detects rising edge points L11 and L13 and falling edge points L12 and L14 from the left side of the captured image, and detects rising edge points R11 and R13 and falling edge points R12 and R14 from the right side of the captured image. To do.

  Then, as shown in FIG. 5, the midpoint calculation unit 22 includes a midpoint between the left falling edge point L14 and the right rising edge point R11, and a midpoint between the left falling edge point L14 and the right rising edge point R13. The middle point of the left rising edge point L13 and the right falling edge point R12, the middle point of the left rising edge point L13 and the right falling edge point R14, the middle point of the left falling edge point L12 and the right rising edge point R11 Point, midpoint of left falling edge point L12 and right rising edge point R13, midpoint of left rising edge point L11 and right rising edge point R12, left rising edge point L11 and right falling edge point R14 Each midpoint is calculated.

  Thereafter, the midpoint position voting unit 23 generates runway candidate units C11 to C18 from the calculated midpoints. Then, after the midpoint position voting unit 23 repeats the voting of the midpoint position and generates a histogram as shown in FIG. 5, the travel path detection unit 24 detects the travel path as described above to detect the travel path. Information is output to the output unit 30.

  Here, when the white line candidate is detected as a line from the edge point as in the conventional case, when the lane mark is a composite line, it is not possible to determine which line is the desired white line, and the estimation result of the traveling path becomes unstable. there were. However, in the road detection device 1 of the present embodiment, as shown in FIG. 5, even when the lane mark is a compound line, the lane mark is symmetric and travels from the middle point of each edge point. Since the road is detected, the estimation result of the travel path is stable and the travel path can be detected accurately.

  Next, an example of a road detection method using the road detection device 1 according to the present embodiment will be described with reference to FIGS. 6 and 7. The process shown in FIG. 6 is a process executed by the ECU 20, and is repeatedly executed, for example, at regular intervals.

First, in step S11 (hereinafter referred to as “S11”, the same applies to other steps), the edge detection unit 21 detects edge points L1 to L4 and R1 from the captured image of the camera 10 as shown in FIG. R4 extraction and labeling are performed for each line (edge detection step). For example, edge points are extracted and labeled along the horizontal line of the road or vehicle in the captured image, and the edge point is extracted along the next line whose position is changed in the vertical direction (traveling direction) of the road or vehicle. Extraction and labeling are performed sequentially. Here, the labeling refers to a process of identifying the extracted edge point group based on the increase / decrease in brightness, height, distance, or the like. The direction of the line may be, for example, the arrangement direction of pixels arranged along the road or the lateral direction of the vehicle in the image to be detected. Next, in S12, a distance (for example, distance d ₁₂ ) between target edges (for example, left edge point L1 and right edge point R2) in the target line (for example, line 1 in FIG. 7) is calculated.

In S13, it is determined whether or not the distance d _ij between the edge points calculated in S12 is greater than the lower limit threshold _Dmin and less than the upper limit threshold _Dmax . Here, the values of the lower limit threshold value D _min and the upper limit threshold value D _max are the lower limit value and the upper limit value that can be assumed as the width of the traveling path of the vehicle, but these values can be changed as appropriate. . If it is determined in S13 that the distance d _ij between the edges is greater than the lower threshold D _min and not less than the upper threshold D _max , the process moves to S15 without voting for the midpoint position. On the other hand, if it is determined in S13 that the distance d _ij between the edges is greater than the lower threshold D _min and less than the upper threshold D _max , the process proceeds to S14, where the midpoint calculator 22 determines the midpoint position between the edges. And the midpoint position voting unit 23 votes the midpoint position (reference point calculation step, reference point position voting step).

  In S15, it is determined whether or not the processes in S12 to S14 have been executed for all edge points of the target line. If it is determined in S15 that the process has been executed for all edge points of the target line, the process proceeds to S16, and if it is determined in S15 that the process has not been executed for all edge points of the target line. The process of S12 to S14 is executed again after changing the target edge.

  In S16, it is determined whether or not the processing in S12 to S14 has been executed for all target lines. If it is determined in S16 that the process has been executed for all target lines, the process proceeds to S18. If it is determined in S16 that the process is not performed for all target lines, the process proceeds to S17.

  In S17, a process for changing the target line to the next line is performed. For example, when the line 1 in FIG. 7 is the target line, the target line is changed to the line 2 positioned below the line 1. In S18, a process of searching for a midpoint position with a large number of votes as a traveling direction of the vehicle is executed. Specifically, the vehicle travels from the vote result of the midpoint position voted by the midpoint position voting unit 23 in S14. A process in which the road detection unit 24 detects a travel path is executed (travel path detection step). As described above, after the travel path detection unit 24 detects the travel path, the travel path information is output to the output unit 30 and the series of processes ends.

  The travel path detection process in the travel path detection apparatus 1 is performed as described above. In the travel path detection apparatus 1, the midpoint position voting unit 23 calculates the midpoint of the rising edge point and the falling edge point as needed. It is possible to increase the number of votes for the midpoint position. Therefore, the estimated midpoint position can be brought closer to the center of the actual lane, and the travel path estimation result can be made more stable. Further, since the road detection device 1 estimates the road by voting the midpoint position, the road detection can be performed even when the lane mark is a branch line or a part of the lane mark is hidden. It can be stabilized.

  Further, in the road detection method according to the first embodiment, the vehicle travels from an edge detection step that detects an edge point that is a change point of the brightness of the road, and the left edge point and the right edge point that are detected in the edge detection step. A reference point calculation step for calculating the reference point, a reference point position voting step for voting the position of the reference point calculated in the reference point calculation step, and a reference point position voted in the reference point position voting step. A travel path detecting step for detecting the travel path. Thus, in the present embodiment, the road is detected using the positions of the reference points obtained from the left and right edge points without detecting the white line candidate as a line as in the prior art. Therefore, even when the actual lane mark is a composite line, the travel path is detected by the voting results of the reference points obtained from the edge points detected from the composite lines on both sides, so the travel path detection result is stable. Can be made.

  Specifically, as shown in FIG. 8 (a), two white lines H21 and H22 extend forward and backward on the left and right sides, and the white line H23 branches further to the left from the left white line H21. In the case of a branch line, the midpoint position is calculated by the midpoint calculator 22 and the midpoint position is voted by the midpoint position voting unit 23 as described above. And it is possible to detect runway candidate portions C21 to C27 suitable as a runway extending forward and backward between the white line H21 and the white line H22.

  That is, as shown in S13 of FIG. 6, it is determined whether the distance between the edge points is within a predetermined range, and the edge point of the branched white line H23 is excluded from the processing target. Suitable runway candidate parts C21-C27 can be detected without receiving. As described above, in the travel path detection device 1, since the travel path is estimated by voting for the midpoint position, the travel path estimation result can be stabilized even when the branched white line H23 is captured.

  Further, as shown in FIG. 8B, the white line H31 extends forward and backward on the left side, the white line H32 extends forward and backward on the right side, and a part of the left white line H31 is hidden by the other vehicle B or the like. In this case, the midpoint of the line hidden by the other vehicle B cannot be detected, but the midpoints of other portions can be detected. Therefore, it is possible to detect runway candidate portions C31 to C35 suitable as a runway extending forward and backward between the white line H31 and the white line H32 before and after the other vehicle B. In this way, even when a part of the white line is hidden, it is possible to avoid a situation in which the estimation result of the midpoint position is significantly shifted, and it is possible to stabilize the estimation result of the traveling road. Furthermore, when the travel route is estimated by voting by the midpoint position voting unit 23, noise resistance is stronger than before, so it is possible to accurately estimate the travel route even in a noisy environment such as rain or at night. It becomes.

(Second Embodiment)
Next, the track detection apparatus 101 of 2nd Embodiment is demonstrated with reference to FIGS. The travel path detection apparatus 101 according to the second embodiment calculates the change amount of the pitching angle of the vehicle, corrects the pitching angle with the change amount, and calculates the distance between the edge points. As shown in FIG. 9, the track detection device 101 according to the second embodiment uses an ECU 120 in which a distance calculation unit 125 and a pitching angle change calculation unit 126 are further added instead of the ECU 20 of the first embodiment. Is different from the first embodiment, and other configurations are the same as those of the first embodiment. Therefore, in the following, the ECU 120 of the second embodiment will be described with emphasis, and description of portions overlapping with those of the first embodiment will be omitted.

The distance calculation unit 125 calculates the distance between the edge point in the left image of the captured image and the edge point in the right image. Specifically, for example, left to calculate the distance d ₁₂ of the rising edge point L1 and a right edge point R2 shown in FIG. 7, the calculation of the distance between such an edge point for each edge point and line-by-line basis Do.

  The pitching angle change calculation unit 126 generates a histogram as shown in FIG. 10 from the distance between the edge points calculated by the distance calculation unit 125 for each edge point and for each line. FIG. 10A shows an example of a half value of the vehicle width when the vehicle does not cause pitching, and FIG. 10B shows an example of a half value of the vehicle width when the vehicle causes pitching. FIG. 10C shows an example in which the pitching angle is corrected by shifting the histogram of FIG. 10B to the right side.

  Here, as shown in FIGS. 10 (a) and 10 (b), the average value X1 of the value obtained by dividing the vehicle width when pitching is not caused by 2 and the vehicle width when pitching is caused. It is different from the average value X2 of the value obtained by dividing by 2. Thus, when pitching occurs, the value of the vehicle width changes on the captured image. Therefore, in this embodiment, the amount of change in the pitching angle is calculated, and after correcting the pitching angle with the amount of change, the distance between the edge points is calculated again, thereby improving the accuracy of detection of the traveling path when pitching occurs. . That is, after correcting the pitching angle, the average value X3 of the frequencies is close to the average value X1 of FIG. 10A as shown in FIG. 10C, which is the same as when no pitching occurs. It is possible to detect the traveling road with the accuracy of

Further, the pitching angle change calculation unit 126 calculates the focal length f of the camera 10, the average value d _av of the vehicle width values of the obtained histogram, and the desired lane width (lane width when the pitching angle is 0 °) d _Lane. , Using a predetermined pitching angle P ₀ , an average value y from the vehicle to the initially set estimation area (an average value of 50 m when the distance from 0 to 100 m ahead of the vehicle is the estimation area), For example, the change amount ΔP of the pitching angle is calculated by the following equation (1).

  Next, a road detection method using the road detection device 101 according to the present embodiment will be described with reference to FIG. The process shown in FIG. 11 is a process executed by the ECU 120, and is repeatedly executed, for example, at regular intervals.

First, in S21, edge point extraction and labeling processing similar to S11 in FIG. 6 is performed. In S22, the distance _dij between the target edge points in the target line is calculated in the same manner as in S12. After S22, a histogram as shown in FIG. 10A using the distance d _ij is created in S23, and an average value d _av or median is calculated from the created histogram (S24).

After S24, it is determined whether or not the absolute value of the difference between the average value d _av or the median value calculated in S24 and the desired lane width d _Lane is less than the threshold value d _thresh (S25). If it is determined in S25 that the absolute value of the difference between the average value d _av or the median value and the lane width d _Lane is less than the threshold value d _thresh , the pitching angle is less than the predetermined value, and the process proceeds to S27. On the other hand, if it is determined in S25 that the absolute value of the difference between the average value d _av or the median value and the lane width d _Lane is not less than the threshold value d _thresh , the pitching angle is assumed to be greater than or equal to a predetermined value, and the process proceeds to S26.

In S26, for example, calculates a change amount [Delta] P of the pitching angle by the equation (1), the value of the pitching angle P as the sum of the predetermined pitching angle P ₀ and the variation amount [Delta] P, the distance d _ij between re edge points calculate. And in S27-S32, the process similar to S13-S18 of 1st Embodiment shown by FIG. 6 is performed, the information of a driving path is output to the output part 30, and a series of processes are complete | finished.

As described above, in the track detection device 101, the pitching angle change calculation unit 126 determines that the difference between the average value d _av or the median value of the distance frequency calculated by the distance calculation unit 125 and the lane width d _Lane that is the width of the road is the threshold value d. If it is equal to or greater than _thresh , the pitching angle change amount ΔP is calculated using the average value d _av or the median value and the lane width d _Lane . When the difference between the average value d _av or the median of the distance frequency and the lane width d _Lane that is the width of the road is equal to or greater than the threshold value d _thresh , the distance calculation unit 125 determines that the midpoint calculation unit 22 Before the point is calculated, the pitch angle P is corrected by the change amount ΔP, and the distance d _ij between the edge points is calculated again.

As described above, in the second embodiment, when pitching occurs, the pitching angle P is corrected and the distance d _ij between the edge points is calculated again. Even when this occurs, the estimation result of the traveling road can be stabilized. Moreover, since the track detection apparatus 101 of 2nd Embodiment is equipped with the structure same as the track detection apparatus 1 of 1st Embodiment, the same effect as the track detection apparatus 1 of 1st Embodiment is also acquired.

  As mentioned above, embodiment mentioned above demonstrates embodiment of the runway detection apparatus which concerns on this invention, and the runway detection apparatus which concerns on this invention is not limited to what was described in the said embodiment. The runway detection apparatus according to the present invention may be modified from the runway detection apparatus according to the above-described embodiment or applied to another so as not to change the gist described in each claim.

  For example, in the above embodiment, the midpoint calculation unit 22 calculates the midpoint of the left edge point and the right edge point, and the midpoint position voting unit 23 votes the midpoint position. A point other than the point may be used as the reference point. That is, instead of setting the position of the reference point as the middle point, for example, the position of the reference point may be set to a position offset by a predetermined amount from the middle point to the left side or the right side. In this case, the vehicle can travel leftward or rightward in the lane. In addition, when the vehicle is traveling on a curve, the position of the reference point can be voted so that the vehicle travels inside the curve in the lane.

  Moreover, in the said embodiment, the edge detection part 21 detected the edge point from the captured image which the camera 10 imaged, and calculated the midpoint between the detected edge points. However, the reference point may be calculated from other than the captured image. That is, instead of the captured image, an infrared sensor or a radar device using electromagnetic waves may be used. For example, the laser device detects a three-dimensional point that is a change point of the brightness of the road, and the reference point is detected from the detected three-dimensional point. May be calculated.

  Moreover, although the example which detects the edge point of a white line was demonstrated in the said embodiment, you may detect the edge point of a curb, for example, and the object which detects an edge point is not limited above.

  Further, in the above embodiment, the edge points are detected from both the left side and the right side of the captured image of the camera 10, but in the case where the information on the travel path on which the vehicle is traveling is acquired, the left side of the captured image is displayed. The edge point may be detected from either one of the right side and the right side, and in this case, the position of the reference point and the traveling path can be estimated.

  Moreover, in the said embodiment, although the edge point was detected from the right-and-left both sides of the road of one lane, the lane of the object road is not limited to one lane. Specifically, for example, the edge point may be detected from a three-lane road by changing the threshold value of the width of the target lane. In this case, even if one white line is hidden, the position of the reference point can be calculated from the white line outside or inside the white line. Therefore, even when the white line is hidden, the traveling path can be detected with high accuracy.

DESCRIPTION OF SYMBOLS 1,101 ... Runway detection apparatus, 21 ... Edge detection part, 22 ... Midpoint calculation part (reference point position calculation part), 23 ... Midpoint position voting part (reference point position voting part), 24 ... Traveling path detection part, 125 ... distance calculation unit, 126 ... pitching angle change calculation unit, d _thresh ... threshold, L1, L3, L11, L13, R1, R3, R11, R13 ... edge points (upward edge points), L2, L4, L12, L14 , R2, R4, R12, R14... Edge point (falling edge point).

Claims (6)

An edge detection unit for detecting an edge point which is a change point of the brightness of the road;
A reference point calculation unit that calculates a reference point that is a candidate for a vehicle traveling path from the left edge point and the right edge point of the road detected by the edge detection unit;
A reference point position voting unit for voting the position of the reference point calculated by the reference point calculation unit;
A travel path detection unit that detects a travel path of the vehicle from the position of the reference point voted by the reference point position voting unit;
A roadway detection device with The reference point is a midpoint between the left edge point and the right edge point.
The runway detection apparatus according to claim 1. The edge detection unit detects rising edge points that increase in brightness when the detection position is moved from one side to the other side on the captured image of the road, and from one side of the captured image to the other side. Detect the falling edge point where the brightness decreases when the position is moved,
The reference point calculation unit calculates the reference point from the left rising edge point and the right falling edge point, and also calculates the reference point from the left falling edge point and the right rising edge point. To calculate,
The runway detection apparatus according to claim 1 or 2. A distance calculator that calculates a distance between the left edge point and the right edge point;
When the difference between the average value or the median value of the distances calculated by the distance calculation unit and the lane width of the road is equal to or greater than a threshold, the average value or the median value and the lane width of the road are used. A pitching angle change calculating unit for calculating a change amount of the pitching angle of the vehicle;
With
When the difference between the average value or the median value and the lane width of the road is equal to or greater than the threshold value, the distance calculation unit performs the pitching before the reference point calculation unit calculates the reference point. Correcting the pitching angle of the vehicle by the amount of change in the pitching angle calculated by the angle change calculating unit, and calculating the distance between the edge points again.
The runway detection device according to any one of claims 1 to 3. The edge point is an edge point of a white line of the road;
The runway detection device according to any one of claims 1 to 4. An edge detection step for detecting an edge point that is a change point of the brightness of the road;
A reference point calculation step of calculating a reference point that is a candidate for a vehicle traveling path from the left edge point and the right edge point on the road detected by the edge detection unit;
A reference point position voting step for voting the position of the reference point calculated in the reference point calculation step;
A travel path detection step of detecting a travel path of the vehicle from the position of the reference point voted in the reference point position voting step;
A runway detection method comprising: