JP4270183B2 - White line detector - Google Patents

Description

  The present invention relates to a white line detection device that detects a white line drawn on a road.

Conventionally, as a white line detection device for detecting a white line drawn on a road, as described in Japanese Patent Application Laid-Open No. 2004-240636, a white line drawn on a driving lane is detected based on a road surface image taken from a vehicle. The image processing apparatus performs edge processing on an image captured in front of the vehicle, detects an edge portion, extracts white line candidate points from the edge portion, and refers to the number or distribution of the white line candidate points. It is known that noise is removed to detect a white line on the road. This apparatus attempts to detect a white line with high accuracy while suppressing an increase in processing amount using only a captured image.
Japanese Patent Laid-Open No. 2004-240636

  However, in such a white line detection device, when there is a shadow of a building or the like on the road surface of the road, it is difficult to detect the white line of the shadowed part, and the white line detection accuracy is lowered. . That is, if there is a shadow on the road surface, the luminance value of the image is lowered overall in the shadow portion, and it is difficult to detect the edge point. Further, if an edge point in a shadow portion is to be detected, noise in a portion other than the shadow portion increases, and white line detection accuracy decreases.

  Therefore, an object of the present invention is to provide a white line detection device capable of suppressing a decrease in white line detection accuracy when there is a shadow on a road.

That is, the white line detection device according to the present invention is a white line detection device that detects a white line drawn on the road based on an image of the road, a shadow region setting unit that sets a shadow region in the image, be one that detects the white line in the image, in the shadow region and a white line detecting means for performing the white line detected as the white line as compared to the area other than the shadow area is likely to be detected, the shadow The region setting means estimates a direction in which the shadow in the image extends in accordance with the position of the sun, and when the direction of the edge line extracted from the image matches the direction in which the shadow extends, the edge line is used as a shadow candidate line. And the shadow region is set based on the shadow candidate line. Here, the white line includes not only a white line but also a non-white line such as yellow, for example, a roadway outer line, a roadway center line, and a lane boundary line.

  In the white line detection device according to the present invention, it is preferable that the white line detection unit increases the contrast of the image in the shadow region so that the white line is easily detected. Also, in the white line detection device according to the present invention, it is preferable that the white line detection means makes the white line easy to detect by lowering a threshold value for white line detection in the shadow region. In the white line detection device according to the present invention, it is preferable that the white line detection unit increases a weight related to an edge in the shadow region so that the white line is easily detected.

According to these inventions, a shadow area is detected in an image, and white line detection is performed in the shadow area so that a white line can be detected more easily than other areas. By detecting the shadow area differently from the other areas in this way, white lines are easily detected in the shadow area. Accordingly, it is possible to suppress a decrease in white line detection accuracy when there is a shadow on the road. Then, by setting the shadow region by estimating the direction in which the shadow extends in the image according to the position of the sun, it is possible to prevent an inappropriate shadow region from being set and to detect an appropriate white line.

  In the white line detection device according to the present invention, when the white line detection unit detects one of the white lines drawn on the left and right sides of the road, the white line detection unit detects the position and angle of the other white line based on the detection result of the one white line. It is preferable to detect at least one of the white lines.

  According to this invention, when one of the white lines on the left and right sides of the road is detected, the white line is detected by estimating at least one of the position and angle of the other white line based on the detection result of the one white line. The position range and angle range of the white line can be narrowed down. For this reason, a white line can be detected with high accuracy. Further, the amount of image processing for detecting the white line can be reduced, and the white line can be detected quickly.

  In the white line detection device according to the present invention, the shadow region setting unit may shadow the intersecting horizontal edge line when the estimated white line set on the image intersects the horizontal edge line extracted from the image. It is preferable to select a candidate line and set the shadow region based on the shadow candidate line.

  According to the present invention, when an estimated white line set on an image and a horizontal edge line extracted from the image intersect, the intersecting horizontal edge line is selected as a shadow candidate line, and a shadow based on the shadow candidate line is selected. By setting the area, it is possible to prevent an inappropriate shadow area from being set and to detect an appropriate white line.

  In the white line detection device according to the present invention, it is preferable that the white line detection unit detects the white line after erasing an edge point of the shadow candidate line.

  According to the present invention, by detecting the white line after erasing the edge point of the shadow candidate line, the edge point other than the edge point of the white line is deleted, so that the detection accuracy of the white line can be improved.

  In the white line detection device according to the present invention, it is preferable that the shadow region is not set when the position of the sun is in a predetermined positional relationship. In the white line detection device according to the present invention, it is preferable that the shadow area is not set when the vehicle is not traveling in an urban area. In the white line detection device according to the present invention, it is preferable not to set the shadow region when the weather is not sunshine.

  According to these inventions, by not setting useless shadow areas, the processing amount required for white line detection can be reduced, and rapid white line detection can be performed.

  ADVANTAGE OF THE INVENTION According to this invention, when a road has a shadow, the fall of the detection accuracy of a white line can be suppressed.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted.

  FIG. 1 is a schematic configuration diagram of a white line detection device according to an embodiment of the present invention.

  As shown in FIG. 1, the white line detection device 1 according to the present embodiment is a device that detects a white line drawn on a road based on an image of the road, and is used by being mounted on a vehicle 5, for example. Here, the white line means a road marking line having a white color or a similar color, and corresponds to, for example, a roadway outer line, a roadway center line, and a lane boundary line. In addition, a color similar to white corresponds to a color that can be displayed as a line, such as a white line, for example, yellow.

  The white line detection device 1 is provided with an imaging unit 2. The imaging unit 2 functions as an imaging unit that images the surroundings of the vehicle, and is installed, for example, so as to capture the front of the vehicle 5. For example, a CCD camera or a C-MOS camera is used as the imaging unit 2.

  The imaging unit 2 is connected to an ECU (Electronic Control Unit) 3. The ECU 3 controls the entire apparatus, and includes, for example, a CPU, a ROM, a RAM, an input signal circuit, an output signal circuit, a power circuit, and the like. The ECU 3 inputs a captured image from the imaging unit 2 and performs image processing to detect a white line drawn on the road.

  Next, the operation of the white line detection device 1 according to this embodiment will be described.

  2 and 3 are flowcharts showing the operation of the white line detection device 1 according to the present embodiment. The series of control processes in FIGS. 2 and 3 are repeatedly executed by the ECU 3 at a predetermined cycle, for example.

  First, as shown in S10 of FIG. 2, an image reading process is performed. The image reading process is a process of reading an image captured by the imaging unit 2 into the ECU 3. As shown in FIG. 4, the read image 31 is an image of the road 20 in front of the vehicle 5. White lines 21 a and 21 b are drawn on the left and right sides of the lane of the road 20. A shadow 22 such as a building extends on the road 20. In addition, when the image read into ECU3 is an analog image, it converts into a digital image by A / D conversion.

  Then, the process proceeds to S12, and edge detection processing is performed. The edge detection process is a process for detecting an edge portion of an image. For example, when a captured image is subjected to edge processing and a change in brightness in the image is greater than a predetermined threshold value, that portion is detected as an edge portion. As shown in FIG. 5, in the edge image 32, one white line 21b is detected as an edge portion, but the portion of the white line 21a covered by the shadow 22 is hardly detected as an edge portion. Further, the outer edge portion of the shadow 22 is detected as an edge portion.

  And it transfers to S14 of FIG. 2 and a noise removal process is performed. The noise removal process is a process for removing noise at the edge portion. For example, processing is performed such that weak edge portions below a predetermined level are removed as noise and only strong edge portions remain. Then, the process proceeds to S16, and white line detection processing is performed. The white line detection process is a process for detecting a white line based on the noise-removed image. For example, a continuous line segment is extracted for the edge portion of the noise-removed image, and a white line is detected based on the extraction result.

  Then, the process proceeds to S18, and it is determined whether or not a white line is detected on only one side of the left and right white lines. If it is determined that the white line is not detected on only one side of the left and right white lines, that is, if both sides of the left and right white lines are detected, the control process ends. On the other hand, if it is determined in S18 that a white line is detected on only one side of the left and right white lines, white line estimation processing is performed (S20). The white line estimation process is a process of estimating the other undetected white line position based on the one white line information detected by the white line detection process. For example, the center of the lane is the center line, and the position of the detected white line and the position of line symmetry are estimated as the position of the undetected white line. In this white line estimation process, the angle of the other undetected white line may be estimated based on one white line information detected by the white line detection process. For example, the center of the lane is the center line, and the angle of the white line that is symmetrical with the detected white line is estimated as the angle of the undetected white line.

  And it transfers to S22 and it is judged whether the weather conditions are satisfied. As the weather condition, a weather condition that shadows the road is set. For example, the weather condition is that the weather is sunny. It is determined that the weather condition is satisfied when the weather is clear, and it is determined that the weather condition is not satisfied when the weather is cloudy, rainy, or snowy. Whether the weather is sunny or not may be determined based on a detection signal from an illuminometer or a rain sensor (not shown), or may be determined based on weather information by external communication.

  Moreover, it is good also as a weather condition that it is the weather which can make a shadow more than predetermined on a road. For example, as shown in FIG. 6, when the position of the sun is within a predetermined angle range with respect to the front-rear direction of the vehicle 5 (for example, within ± 15 degrees, see the hatched portion in FIG. 6), Therefore, it is determined that the weather does not cause a shadow of a predetermined level or more, and it is determined that the weather condition is not satisfied. On the other hand, when the position of the sun is not within a predetermined angle range with respect to the front-rear direction of the vehicle 5, it is determined that the weather is more than a predetermined shadow on the road because it is easy to make a shadow across the road, It is determined that the weather conditions are met. Whether the position of the sun is within a predetermined angle range with respect to the front-rear direction of the vehicle 5 is, for example, information on the position of the sun calculated from time information including the date and vehicle traveling direction information from a navigation system or a gyro Judgment may be made based on the above.

  In addition, as shown in FIG. 7, when the position of the sun is within a predetermined angle range with respect to the vertical direction of the vehicle 5 (for example, within ± 30 degrees; see the hatched portion in FIG. 7), a long shadow on the road Therefore, it is determined that the weather is not capable of shadowing more than a predetermined amount on the road, and it is determined that the weather condition is not satisfied. On the other hand, when the position of the sun is not within a predetermined angle range with respect to the vertical direction of the vehicle 5, a long shadow is likely to be formed on the road. Is determined to be true. Whether or not the position of the sun is within a predetermined angle range with respect to the vertical direction of the vehicle 5 may be determined based on, for example, position information of the sun calculated from time information including the date.

  If it is determined in S22 of FIG. 2 that the weather condition is not satisfied, the control process is terminated. On the other hand, if it is determined in S22 that the weather condition is satisfied, horizontal edge detection processing is performed (S24). The horizontal edge process is a process for detecting a horizontal edge portion of an image. For example, horizontal edge processing is performed on the captured image, and when the change in brightness in the horizontal direction of the image is greater than a predetermined luminance value, that location is detected as a horizontal edge portion. Then, the process proceeds to S26, and edge value narrowing processing is performed. This edge value narrowing-down process is a process of removing a horizontal edge portion having an edge value smaller than a preset narrowing-down threshold value.

  Then, the process proceeds to S28 in FIG. 3 and a Hough conversion process is performed. The Hough transform process is a process for detecting a horizontal edge line segment using the Hough transform. By this processing, horizontal edge points having a line segment relationship are detected as line segments. Then, the process proceeds to S30, and it is determined whether or not the horizontal edge line segment detected by the Hough transform process intersects with a division line set in advance in the image. For example, as shown in FIG. 8, in the horizontal edge image 33 that has been subjected to the Hough transform process, the division lines 34a and 34b that extend perpendicularly to the area where the white line is to be detected are set. The dividing lines 34a and 34b are set one by one at the left and right positions of the road. In S30, it is determined whether or not the horizontal edge line segment 35 intersects the dividing line 34a on the side where the white line is not detected in S18. In FIG. 8, the horizontal edge line segment 35 intersects the dividing line 34a on the side where the white line is not detected.

  If it is determined in S30 that the horizontal edge line segment does not intersect the dividing line set in the image, the control process is terminated. On the other hand, when it is determined in S30 that the horizontal edge line segment intersects the dividing line set in the image, shadow direction estimation processing is performed (S32). This shadow direction estimation process is a process of estimating the direction in which the shadow extends according to the position of the sun. For example, the direction in which the shadow formed on the road extends is estimated based on the position information of the sun and the traveling direction information of the vehicle.

  Then, the process proceeds to S34, and shadow candidate line extraction processing is performed. This shadow candidate line extraction process is a process of extracting a line segment that matches the estimated shadow direction as a shadow candidate line in the horizontal edge image. For example, in the horizontal edge image 33 of FIG. 8, an edge line segment 35 that matches the shadow direction A is extracted as a shadow candidate line. At this time, the case of matching with the shadow direction A includes not only the case of complete matching but also the case of substantially matching (for example, the case where the angle deviates from the shadow direction A by several degrees or 10 degrees).

  Then, the process proceeds to S36, and shadow area setting processing is performed. The shadow area setting process is a process of setting a shadow area on the road of the image based on the shadow candidate line. For example, a predetermined range is set as a shadow area centering on the position of the shadow candidate line. Then, the process proceeds to S38, and edge detection processing in the shadow area is performed. This edge detection process is a process for detecting an edge in a shadow area.

  In this edge detection process, white lines are detected so that white lines are more easily detected than in areas other than shadow areas. For example, the white line can be easily detected by lowering the threshold for white line detection. Specifically, the edge detection may be performed using a threshold value smaller than the threshold value used in the edge detection process of S12. This makes it possible to detect a white line with a small change in luminance value in the shadow area.

  In this shadow area edge detection process, the contrast of the image in the shadow area may be increased so that white lines can be easily detected. For example, by performing edge detection with the luminance value of each pixel in the shadow region being doubled or more, it is possible to detect an edge with a small change in luminance value, and to easily detect a white line.

  Further, in the edge detection process of the shadow area, the weight related to the edge in the shadow area may be increased so that the white line can be easily detected. For example, when a line segment is detected by Hough transform or the like for an edge detected in a shadow area, a white line is easily detected in the shadow area by increasing the weight with respect to another.

  Then, the process proceeds to S40, and edge point removal processing is performed. The edge point removal process is a process for removing edge points on the shadow candidate line. By this process, an edge point formed by a shadow on the road can be removed. Then, the process proceeds to S42 and a Hough conversion process is performed. This Hough transform process is a process for detecting a line segment of an edge portion of a shadow region using the Hough transform. In that case, it is preferable to detect a line segment using the white line position estimated in S20.

  Then, the process proceeds to S42, and white line redetection processing is performed. This white line redetection process is a process of performing white line detection on the entire image using edge line segments detected in the shadow area. When this white line redetection process is completed, a series of control processes are terminated.

  As described above, according to the white line detection device 1 according to the present embodiment, a shadow area is detected in a captured image, and a white line is detected in the shadow area so that a white line can be detected more easily than other areas. I do. Thus, by detecting the white line different from the other areas in the shadow area, the white line is easily detected in the shadow area. Accordingly, it is possible to suppress a decrease in white line detection accuracy when there is a shadow on the road.

  Moreover, according to the white line detection device 1 according to the present embodiment, when one of the white lines on the left and right sides of the road is detected, at least one of the position and the angle of the other white line is estimated based on the detection result of the one white line. By detecting the white line, the position range and angle range of the other white line can be narrowed down. For this reason, a white line can be detected with high accuracy. Further, the amount of image processing for detecting the white line can be reduced, and the white line can be detected quickly.

  Further, according to the white line detection device 1 according to the present embodiment, setting the shadow area by estimating the direction in which the shadow extends in the image according to the position of the sun prevents the setting of an inappropriate shadow area. Appropriate white line detection can be performed.

  Also, when the estimated white line set on the image and the horizontal edge line extracted from the image intersect, select the intersecting horizontal edge line as the shadow candidate line and set the shadow area based on the shadow candidate line By doing so, it is possible to prevent an inappropriate shadow area from being set and to detect an appropriate white line.

  In addition, according to the white line detection device 1 according to the present embodiment, the white line detection is performed after the edge point of the shadow candidate line is erased, so that the edge points other than the white line edge point are erased. The accuracy can be improved.

  In addition, when the sun position is in a predetermined positional relationship, the shadow area is not set, so that it is not necessary to set a useless shadow area, and the processing amount required for white line detection can be reduced, and promptly. White line detection can be performed. Also, when the weather is not sunshine, the shadow area is not set, so it is not necessary to set a useless shadow area, the amount of processing required for detecting the white line can be reduced, and the white line can be detected quickly. .

  The shadow area may not be set when the vehicle is not traveling in an urban area. For example, after S22 in FIG. 2, a determination process is performed to determine whether or not the vehicle is traveling in an urban area. If the vehicle is traveling in an urban area, the process proceeds to S24, and the vehicle is not traveling in an urban area. The control process ends. Whether or not the vehicle is traveling in an urban area may be determined based on the map information of the navigation system. At that time, if it is within a predetermined distance from the city area, it may be determined that the city area is traveling. The shadow area may not be set according to the road on which the vehicle is traveling. In these cases, setting of useless shadow areas can be avoided, the amount of processing required for white line detection can be reduced, and rapid white line detection can be performed.

  The embodiment described above shows an example of the white line detection device according to the present invention. The white line detection device according to the present invention is not limited to the above, and the white line detection device according to the embodiment is modified or applied to others so as not to change the gist described in each claim. It may be a thing.

1 is a schematic configuration diagram of a white line detection device according to an embodiment of the present invention. It is a flowchart which shows the white line detection process in the white line detection apparatus of FIG. It is a flowchart which shows the white line detection process in the white line detection apparatus of FIG. It is explanatory drawing of the captured image in the white line detection apparatus of FIG. It is explanatory drawing of the image which edge-processed in the white line detection apparatus of FIG. It is explanatory drawing of the weather condition judgment process in the white line detection apparatus of FIG. It is explanatory drawing of the weather condition judgment process in the white line detection apparatus of FIG. It is explanatory drawing of the image which carried out the horizontal edge process in the white line detection apparatus of FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... White line detection apparatus, 2 ... Imaging part, 3 ... ECU, 5 ... Vehicle, 20 ... Road.

Claims (10)

In the white line detection device for detecting a white line drawn on the road based on an image of the road,
A shadow area setting means for setting a shadow area in the image;
White line detection means for detecting the white line in the image, the white line detection means for detecting the white line so that the white line is more easily detected in the shadow area compared to an area other than the shadow area;
With
The shadow region setting means estimates a direction in which the shadow extends in the image according to the position of the sun, and if the direction of the edge line extracted from the image matches the direction in which the shadow extends, A candidate line and setting the shadow region based on the shadow candidate line;
A white line detection device characterized by the above. The shadow area setting means, when an estimated white line set on the image and a horizontal edge line extracted from the image intersect, selects the intersecting horizontal edge line as a shadow candidate line, and the shadow candidate line Setting the shadow area based on
The white line detection device according to claim 1. 3. The white line detection device according to claim 1, wherein the white line detection unit detects the white line after erasing an edge point of the shadow candidate line. 4. The white line detection device according to claim 1, wherein the shadow region is not set when the position of the sun is in a predetermined positional relationship. The white line detection device according to claim 1, wherein the shadow region is not set when the vehicle is not traveling in an urban area. The white line detection device according to any one of claims 1 to 5, wherein the shadow region is not set when the weather is not sunshine. The white line detection device according to claim 1, wherein the white line detection unit increases the contrast of an image in the shadow region so that the white line is easily detected. The white line detection device according to claim 1, wherein the white line detection unit makes the white line easy to be detected by lowering a threshold value for white line detection in the shadow region. The white line detection device according to claim 1, wherein the white line detection unit increases a weight related to an edge in the shadow region so that the white line is easily detected. The white line detecting means detects a white line by estimating at least one of the position and the angle of the other white line based on a detection result of the white line when one of white lines drawn on the left and right sides of the road is detected. thing,
The white line detection device according to any one of claims 1 and 7 to 10.

Images