WO2016189777A1

WO2016189777A1 - Road surface marking detection device, obstruction element detection device, traffic lane detection device, traffic lane detection method, and program

Info

Publication number: WO2016189777A1
Application number: PCT/JP2016/001168
Authority: WO
Inventors: 成俊鴇田
Original assignee: 株式会社Ｊｖｃケンウッド
Priority date: 2015-05-28
Filing date: 2016-03-03
Publication date: 2016-12-01
Also published as: JP2016224585A

Abstract

Provided is a road surface marking detection device, comprising a road surface marking detection unit (22) which: segments an edge image, which is generated on the basis of a captured image of a road, into left and right blocks at the approximate center of the edge image; extracts, from prescribed pixel rows of the edge image, the pixel rows in which the distances between the pixels which are nearest to the approximate center in both the left and right blocks of the pixel rows are less than or equal to a prescribed value; and determines whether a road surface marking is present on the basis of the percentage of the number of extracted pixel rows with respect to the number of the prescribed pixel rows. It is thus possible to detect, with a simple configuration, a crosswalk or other road surface marking in toto.

Description

Road marking detection device, obstacle element detection device, lane detection device, lane detection method and program

The present invention relates to a road marking detection device, an obstacle element detection device, a lane detection device, a lane detection method, and a program.

A lane detection device that detects lane boundary lines and the like has been developed to determine the position of the host vehicle. For example, Patent Document 1 describes that the lane detection device performs a process of extracting a white line (lane boundary line) and a process of extracting a road marking such as a pedestrian crossing.

JP 2011-154480 A

In the lane detection device according to the background art, after extracting a white line candidate, a pedestrian crossing, an arrow sign, and a speed sign are individually extracted and excluded from the white line candidate, and a white line is determined. In other words, this lane detection device is a high-speed and expensive device with an increased image processing amount in order to individually extract road markings such as pedestrian crossings.

The present invention has been made to solve such problems, and can detect road markings such as a pedestrian crossing together with a simple configuration or method, a road marking detection apparatus, an obstacle element detection apparatus, a lane It is an object to provide a detection device, a lane detection method, and a program.

Therefore, in this embodiment, an edge image generated based on an image of a road is divided into left and right blocks at the approximate center of the edge image, and a predetermined pixel row of the edge image is divided into left and right blocks in the pixel row. A pixel row in which the distance between the pixels closest to the approximate center of each of the blocks is less than or equal to a predetermined value is extracted, and the road surface is based on the ratio of the number of extracted pixel rows to the number of the predetermined pixel rows. Provided is a road marking detection device including a road marking detection unit that determines the presence or absence of a sign.

Further, in the present embodiment, an edge image generated based on an image of a road is divided into left and right blocks at the approximate center of the edge image, and a predetermined pixel row of the edge image is divided into right and left in the pixel row. A pixel row in which the distance between the pixels closest to the approximate center of each of the blocks is less than or equal to a predetermined value, and the number of extracted pixel rows is based on a ratio of the number of the predetermined pixel rows Provided is a failure element detection device including a failure element detection unit that determines the presence or absence of a failure element including a vehicle.

In the present embodiment, a lane boundary detection unit that detects a lane boundary candidate from an edge image generated based on an image obtained by imaging a road, and the edge image is divided into left and right blocks at substantially the center of the edge image. Dividing and extracting, from the predetermined pixel row of the edge image, a pixel row whose distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less, The road marking detection unit that determines the presence or absence of road marking based on the ratio of the number to the number of the predetermined pixel rows, and the lane boundary detection unit when the road marking detection unit determines that there is no road marking. There is provided a lane detection device including a determination unit that determines a lane boundary candidate detected by the vehicle as a lane boundary.

Further, in the present embodiment, a step of detecting a lane boundary line candidate from an edge image generated based on an image obtained by imaging a road, and a step of dividing the edge image into left and right blocks at substantially the center of the edge image; Extracting from the predetermined pixel row of the edge image a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is equal to or less than a predetermined value; and The step of determining the presence or absence of road marking based on the ratio of the number to the number of the predetermined pixel rows, and determining that there is no road marking, the detected lane boundary line candidate is determined as the lane boundary line And a lane detection method.

In the present embodiment, the computer detects a lane boundary line candidate from an edge image generated based on an image of a road, and divides the edge image into left and right blocks at the approximate center of the edge image. And a procedure for extracting, from a predetermined pixel row of the edge image, a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less, and Based on the ratio of the number of pixel rows to the number of the predetermined pixel rows, the procedure for judging the presence or absence of road markings, and the detected lane boundary line candidates when it is judged that there are no road markings, And a program for executing the procedure for judging.

According to the present embodiment, it is possible to provide a road marking detection device, an obstacle element detection device, a lane detection device, a lane detection method, and a program that collectively detect road markings such as pedestrian crossings with a simple configuration or method.

1 is a block diagram showing a schematic configuration of a lane detection system 1 according to an embodiment. It is a flowchart which shows the process sequence of the lane detection method which concerns on embodiment. It is a flowchart which shows the detailed process sequence of the center side edge image generation process which concerns on embodiment. It is a flowchart which shows the detailed process sequence of the road marking detection process which concerns on embodiment. It is an example of a bird view image, a thin line edge image, and a center side edge image when there is no road marking according to the embodiment. It is an example of a bird view image, a thin line edge image, and a center side edge image when the road marking according to the embodiment is a pedestrian crossing. It is an example of a bird view image, a thin line edge image, and a center side edge image when the road marking according to the embodiment is the maximum speed. It is an example of a bird view image, a thin line edge image, and a center side edge image when the road marking according to the embodiment has a pedestrian crossing or a bicycle crossing zone. It is an example of the bird view image, thinning edge image, and center side edge image in case the road marking which concerns on embodiment is a place name. It is an example of the bird view image, thinning edge image, and center side edge image when the road marking according to the embodiment is the traveling direction. It is a figure which shows the example of the process result of the road marking detection process which concerns on embodiment. It is a flowchart which shows the detailed process sequence of the lane determination process which concerns on embodiment.

Hereinafter, a lane detection system according to the present embodiment will be described with reference to the drawings.
In addition, in this specification, in order to make the explanation easy to understand, it is called a road center line, a lane boundary line, a road outer line, etc. In contrast to this, there is a pedestrian crossing, maximum speed, pedestrian crossing or bicycle crossing zone, place name, direction of travel, etc., and what is drawn in white on the road surface is collectively called road marking I will decide. In the present specification, the above-described lane boundary line is not included in the road marking.

For the names and shapes of lane boundaries and road markings, refer to the following electronic technical information.
Ministry of Land, Infrastructure, Transport and Tourism, “System of Road Technical Standards”, [online], [Search May 28, 2015], Internet <http://www.mlit.go.jp/road/sign/kijyun/taikei01.html >

First, the configuration of the lane detection system according to the present embodiment will be described.
FIG. 1 is a block diagram showing a schematic configuration of a lane detection system 1 according to the present embodiment.
The lane detection system 1 includes a camera 10, a lane detection unit 20, a monitor 30, a speaker 40, and the like.
The camera 10 is an in-vehicle camera that captures a road in the traveling direction of the host vehicle, that is, in front of or behind the host vehicle.

The lane detection unit 20 performs various image processing on the image input from the camera 10 and detects a lane in which the host vehicle travels, that is, a travel lane. Further, the lane detection unit 20 outputs a warning signal to the monitor 30 and the speaker 40 when the host vehicle deviates from the lane.
The monitor 30 inputs a warning signal and outputs a warning display, and the speaker 40 inputs the warning signal and reproduces a warning sound, and notifies each driver of the danger.

The lane detection unit 20 includes a lane boundary detection unit 21, a road marking detection unit 22, a determination unit 23, and the like.
The lane boundary detection unit 21 performs edge extraction, Hough conversion, etc., extracts straight lines, and detects lane boundary candidates.
The road marking detection unit 22 collectively detects road markings such as a pedestrian crossing, a maximum speed, a pedestrian crossing or a bicycle crossing zone.
The determination unit 23 determines a lane based on the detection results of the lane boundary detection unit 21 and the road marking detection unit 22.
Specific operation of the lane detector 20 will be described later.

In addition, each component which the lane detection part 20 implement | achieves is realizable by making a program run by control of the arithmetic unit (not shown) with which the lane detection part 20 which is a computer is provided, for example. More specifically, the lane detection unit 20 is realized by loading a program stored in a storage unit (not shown) into a main storage device (not shown) and executing the program under the control of the arithmetic unit. Each constituent element is not limited to being realized by software by a program, but may be realized by any combination of hardware, firmware, and software.

The program described above can be stored using various types of non-transitory computer readable media and supplied to a computer. Non-transitory computer readable media include various types of tangible storage media (tangible storage medium). Examples of non-transitory computer-readable media include magnetic recording media (eg flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (eg magneto-optical discs), CD-ROMs (Read Only Memory), CD-Rs, CD-R / W, semiconductor memory (for example, mask ROM, PROM (Programmable ROM), EPROM (Erasable ROM), flash ROM, RAM (random access memory)) are included.

Also, the program may be supplied to a computer by various types of temporary computer readable media. Examples of transitory computer readable media include electrical signals, optical signals and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire and an optical fiber, or a wireless communication path.

Then, operation | movement of the lane detection system 1 which concerns on this Embodiment is demonstrated. Here, the operation of the lane detection unit 20, that is, the lane detection method will be particularly described.
FIG. 2 is a flowchart showing a processing procedure of the lane detection method according to the present embodiment.
Of the processes related to steps S10 to S150 shown in FIG. 2, the processes related to steps S10 to S90 are mainly executed by the lane boundary detection unit 21, and the processes related to steps S100 to S110 are mainly performed on the road surface. The sign detection unit 22 executes and the processes related to steps S120 to S150 are mainly executed by the determination unit 23, but the sharing of these processes is not limited to the above.

First, when the lane detector 20 starts operating, the lane detector 20 inputs a color image in front of the host vehicle photographed by the camera 10 (step S10).
Then, the input color image is converted into a grayscale image having only luminance information (step S20). Of course, it is also possible to input a grayscale image from the camera 10 and omit step S20.

Then, a process of resizing the grayscale image to reduce the amount of computation, for example, an image of horizontal 640 pixels × vertical 480 pixels (x direction 640 pixels × y direction 480 pixels) is converted into an image of 320 pixels vertical × 240 pixels horizontal. Reduction processing is performed, and the resized image is converted into a bird view image (a bird's-eye view image looking down from directly above) (step S30).
Then, noise reduction processing, such as Gaussian filter processing, is performed on the bird view image (step S40). This process has the effect of facilitating extraction of edges as well as the effect of noise reduction.

Then, edge extraction processing is performed on the noise-reduced image using a Sobel filter or the like (step S50). Generally, since the brightness difference between the white line of the lane boundary line and the asphalt surface is large, the boundary (edge) between the lane boundary line and the asphalt surface can be appropriately extracted by this processing.
In the grayscale edge image obtained by the edge extraction process in step S50, an edge with a large luminance difference is a color close to white (luminance: 255), and an edge with a small luminance difference is a color close to black (luminance: 0). ing. Here, binarization processing is performed in order to leave only edges with a large luminance difference (step S60). By binarization processing, a pixel having a luminance value larger than a predetermined threshold is set to white, a pixel having a luminance value lower than the threshold is set to black, and a binary image (monochrome image) having only a large luminance difference is obtained. ) Is obtained.

Then, the obtained binarized image is thinned, and the thick line is thinned to obtain a thinned edge image (step S70). In this specification, if the thinned edge image is, for example, an image of 320 × 240 pixels, the coordinates of the upper left pixel of the image are (x, y) = (0, 0), and the coordinates of the lower right pixel are (x , Y) = (319,239).
And the process which detects a lane boundary line candidate is performed on the one hand using the thinned edge image obtained, and the process which detects a road marking is performed on the other hand.

When detecting a lane boundary line candidate, first, a Hough conversion process is performed on the obtained thinned edge image to extract a straight line (step S80). The extracted straight line may include not only the edge of the lane boundary line but also the edge of the preceding vehicle or guardrail, the edge of the road marking, and the like. Therefore, the lane boundary line is “the straight line is almost vertical (substantially parallel to the vertical direction of the image)”, “the distance between the left straight line and the right straight line is the road width on the road surface (3-4 m The left lane boundary line candidate and the left lane boundary candidate from among the extracted straight lines using features such as “corresponding to the degree)” and “continuous in time (similar positions in consecutive frames)” The right lane boundary line candidate is detected as a set (pair) (step S90).

When detecting a road marking, a center-side edge image is generated from the thinned edge image obtained in step S70 (step S100), and the road marking is detected from the center-side edge image (step S110).
FIG. 3 is a flowchart showing a detailed processing procedure of the center edge image generation processing according to the present embodiment. The center-side edge image generation process (step S100) is specifically composed of steps S101 to S104. Step S100 is not followed by steps S101 to S104. By generating the edge image near the center, it becomes easy to detect the road marking from the thinned edge image.

First, the obtained thinned edge image is divided into a left block and a right block at the approximate center of the image (step S101). For example, if the thinned edge image is an image of 320 × 240 pixels, the left block and the right block are each a block of 160 × 240 pixels.
Then, for a certain y coordinate (a certain pixel row), it is determined whether there are a plurality of edges in the left block or the right block (steps S102 and S103).

If there are a plurality of edges in the left block or the right block (Yes in step S103), one pixel closer to the center of the plurality of edges, for example, if the thinned edge image is an image of 320 × 240 pixels In the left block, one pixel closest to the approximate center x = 159 is left, and in the right block, one pixel closest to the approximate center x = 160 is left, the remaining pixels are deleted (step S104), and the process returns to step S102.
If the left block and the right block do not have a plurality of edges (No in step S103), the process directly returns to step S102.

The same process (steps S102 to S104) is performed for the next y coordinate (next pixel row).
Then, the above processing is performed for all the y coordinates (for example, if the left block and the right block are blocks of 160 × 240 pixels, y = 0, 1, 2,... 238, 239). (Steps S102 to S104) are performed to generate an edge image near the center.
If there is no next y coordinate (No in step S102), the center edge image generation process (step S100) is terminated, and the process proceeds to the road marking detection process (step S110).

FIG. 4 is a flowchart showing a detailed processing procedure of road marking detection processing according to the present embodiment. The road marking detection process (step S110) includes steps S111 to S118 in detail.
First, it is determined whether there is an edge in both the left and right blocks with respect to a certain y coordinate (a certain pixel row) of the center-side edge image (step S111: Yes, step S112).

If there is an edge in both the left and right blocks (Yes in step S112), it is determined whether the edge interval is equal to or smaller than a predetermined value (or less than a predetermined value) (step S113). As this predetermined value, a value corresponding to an assumed lane width (travel lane width, 3 to 4 m) of the road surface is used.
If the edge interval is less than or equal to the predetermined value (or less than the predetermined value) (Yes in step S113), 1 is added to the road marking counter value (step S114), and the process returns to step S111.

If both the left and right blocks do not have an edge (No in step S112), or if the edge interval is not less than a predetermined value (or less than the predetermined value) (No in step S113), the process returns to step S111 as it is. .
The same processing (steps S111 to S114) is performed for the next y coordinate (next pixel row) of the center edge image.

Then, the above processing (steps S111 to S114) is performed for all y coordinates (all pixel rows) of the center edge image. If there is no next y coordinate (No in step S111), road marking The rate is calculated (step S115). The road marking rate is defined by the ratio of the road marking counter value to the height of the center edge image (for example, 240 if the center edge image is an image of 320 × 240 pixels). If the road marking rate is greater than or equal to a certain threshold value, it can be determined that there is a high possibility that a road marking exists.

FIG. 5 is an example of a bird view image, a thinned edge image, and a center edge image when there is no road marking according to the present embodiment. In order from the left, the image after bird view conversion in step S30, the thinned edge image after dividing into the left and right blocks in step S101, and the center edge image after the road marking rate calculation in step S115 are arranged.
The bird view image shows two lanes that gently curve to the left and the roadside band on the left, the preceding vehicle in the left lane, two vehicles in the right lane, and the left side of the roadside band Each guardrail is shown. The vehicle is on the left lane, the left lane boundary is partially cut off, and the right lane boundary is largely interrupted.

In the thinned edge image, there is a dotted line that divides the thinned edge image into left and right blocks in the center vertical direction, and in order from the left, the guardrail, the left and right lane boundary lines, and the edge of the vehicle in the right lane Is reflected.
In the center edge image, dotted lines that are divided into left and right blocks are similarly shown, and the edges of the left and right lane boundary lines and a part of the right lane of the vehicle are shown without being deleted.
Further, the road marking counter value and the road marking rate (inside []) of this image are shown in the lower left corner of the center-side edge image, and the road marking counter value when there is no road marking is 0. The rate is also 0%.

FIG. 6 is an example of a bird view image, a thinned edge image, and a central edge image when the road marking according to the present embodiment is a pedestrian crossing. The dotted lines for dividing the thinned edge image and the center-side edge image into the left and right blocks are omitted (the same dotted lines are also omitted in FIGS. 7 to 10).
A pedestrian crossing is shown in the lower center of the bird view image, and an edge of the pedestrian crossing is shown in the lower center of the thinned edge image. In addition, in the edge image close to the center, when the edge interval is equal to or smaller than a predetermined value (Yes in step S113), the space between the edges is filled (the edges are also painted similarly in FIGS. 7 to 10). When the road marking is a pedestrian crossing, the road marking counter value is 34 and the road marking rate is 14%.

FIG. 7 is an example of a bird view image, a thinned edge image, and a center edge image when the road marking according to the present embodiment is at the maximum speed.
The maximum speed “40” is displayed at the center of the bird view image, and the lane boundary line is displayed to the right of the maximum speed. When the road marking is at the maximum speed, the road marking counter value is 91 and the road marking rate is 37%.

FIG. 8 is an example of a bird view image, a thinned edge image, and a center edge image when the road marking according to the present embodiment has a pedestrian crossing or a bicycle crossing zone.
A road marking with a pedestrian crossing or a bicycle crossing zone is displayed at the lower center of the bird view image, and a lane boundary line is shown on the left and right. When the road marking is a pedestrian crossing or a bicycle crossing zone, the road marking counter value is 67 and the road marking rate is 27%.

FIG. 9 is an example of a bird view image, a thinned edge image, and a center edge image when the road marking according to the present embodiment is a place name.
The place names “Omiya” and “Ginza” appear in the center of the bird view image, and multiple lane boundaries appear on the left and right. When the road marking is a place name, the road marking counter value is 114, and the road marking rate is 47%.

FIG. 10 is an example of a bird view image, a thinned edge image, and a center edge image when the road marking according to the present embodiment is in the traveling direction.
In the lower part of the bird view image, arrows indicating the direction of travel “left turn”, “straight or left turn”, and “right turn” are displayed in order from the left. When the road marking is in the traveling direction, the road marking counter value is 63 and the road marking rate is 26%.

FIG. 11 is a diagram showing an example of the result of the road marking detection process according to the present embodiment. FIG. 11 summarizes the processing results of the images shown in FIGS.
Thus, when the road marking rate is equal to or higher than a certain threshold value, it can be determined that there is a high possibility that the road marking exists.

Therefore, it is determined whether the road marking rate calculated in step S115 is equal to or greater than a predetermined threshold (for example, 10%) (step S116). If the road marking rate is greater than or equal to a predetermined threshold (Yes in step S116), it is determined that a road marking has been detected or there is a road marking (step S117). If the road marking rate is not equal to or greater than the predetermined threshold (No in step S116), it is determined that no road marking has been detected or there is no road marking (step S118).
Then, the road marking detection process (step S110) ends.

For the thinned edge image generated in step S70, the left lane boundary candidate and the right lane boundary candidate are detected on the one hand (step S90), and the road marking is detected on the other hand (step S110). (Own vehicle travel lane) is determined (step S120).
FIG. 12 is a flowchart showing a detailed processing procedure of the lane determination processing according to the present embodiment. Specifically, the lane determination process (step S120) includes steps S121 to S123.

First, it is determined whether a road marking has been detected in step S110 (step S121).
When the road marking is not detected (No in Step S121), the lane is determined based on the left and right lane boundary line candidates detected in Step S90 (Step S122).
If a road marking is detected (Yes in step S121), the left and right lane boundary line candidates detected in step S90 are highly likely to be erroneously detected, so the lanes up to the previous frame (host vehicle lane) (Step S123) to reduce malfunctions.
Then, the lane determination process (step S120) ends.

Then, based on the lane determined in step S120, it is determined whether the host vehicle deviates from the lane (step S130). The determination method here may be a known method.
If the host vehicle deviates from the lane (Yes in step S130), the warning signal is output as described above (step S140), and it is determined whether the process is continued (step S150). If the host vehicle does not depart from the lane (No in step S130), it is determined whether or not to continue the process (step S150).
When the process is continued (Yes in step S150), the process returns to step S10. When the process is not continued (No in step S150), the vehicle detection method is terminated.

In the lane detection method according to the present embodiment, when the road marking is detected (Yes in step S121), the lane up to the previous frame is held (step S123), but when the road marking is detected, A lane boundary line may be selected from left and right lane boundary line candidates close to the position of the lane boundary line up to the previous frame.

In the lane detection method according to the present embodiment, each process is performed using all the y coordinates (all pixel rows) of the image input from the camera 10. Each process may be performed using some pixel rows). For example, in the center edge image generation process (step S100), the road marking detection process (step S110), etc., y coordinates (pixels) that are part of the thinned edge image generated in step S70 or that are separated at a predetermined interval. (Line) may be used for processing. The same applies to the x coordinate (pixel array) of the image input from the camera 10.

In the lane detection method according to the present embodiment, the lane boundary detection unit detects the left and right lane boundary candidates in pairs (pairs), but the lane boundary detection unit detects one of the left and right lane boundary candidates. Then, the lane may be detected. For example, the left and right lane boundary lines are disappearing (or disappearing) and only one lane boundary line candidate can be detected. At this time, the determination unit detects the lane by estimating the position of the other lane boundary line from the detected one lane boundary line candidate. This is based on the fact that the lane width does not change suddenly and is often 3-4 m. The determination unit stops the process of estimating the position of the other lane boundary line when the road marking is detected, and the other lane when the road marking is not detected. You may perform the process which estimates the position of a boundary line.

In the lane detection method according to the present embodiment, an in-vehicle camera is used as the camera 10. However, a camera other than the in-vehicle camera, for example, a camera that can shoot from the sky may be used. In this case, the bird view conversion process (step S30) of the input image is unnecessary.
Similarly, the resizing process (step S30), the noise reduction process (step S40), the binarization process (step S60), the thinning process (step S70), the center edge image generation process (step S100), and the like are also performed. It can be omitted in consideration of detection accuracy and processing speed of line candidates and road markings.

In the lane detection method according to the present embodiment, when the inter-vehicle distance from the preceding vehicle is small, edges such as the rear wiper of the preceding vehicle may gather at the center of the image and the road marking rate may increase. By using this, the lane detection method according to the present embodiment is configured as a preceding vehicle detection method, a failure element detection method for detecting a failure element that may cause erroneous lane detection such as road markings and preceding vehicles, and the like. May be.
Similarly, the lane detection system according to the present embodiment may be configured as a road marking detection device, a lane detection device, a lane detection method, a lane detection program, and the like.

As described above, the road marking detection apparatus according to the present embodiment divides an edge image generated based on an image obtained by imaging a road into right and left blocks at a substantially center of the edge image, and performs predetermined processing on the edge image. A pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is less than or equal to a predetermined value from the pixel row, and the number of extracted pixel rows is the number of the predetermined pixel row A road marking detection unit 22 that determines the presence / absence of road marking based on the ratio of the number to the number is provided.
With such a configuration, road markings such as pedestrian crossings can be collectively detected with a simple configuration.

In addition, the obstacle element detection device according to the present embodiment divides an edge image generated based on an image obtained by capturing a road into right and left blocks at the approximate center of the edge image, and starts from a predetermined pixel row of the edge image. A pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is less than or equal to a predetermined value is extracted, and the number of extracted pixel rows occupies the number of the predetermined pixel rows A failure element detection unit that determines the presence or absence of a failure element including a preceding vehicle based on the ratio is provided.
With such a configuration, it is possible to easily detect an obstacle element including a preceding vehicle with a simple configuration.

Further, the lane detection device according to the present embodiment includes a lane boundary detection unit 21 that detects a lane boundary candidate from an edge image generated based on an image obtained by capturing a road, and the edge image is an abbreviation of the edge image. Divide into left and right blocks at the center, and extract from the predetermined pixel row of the edge image a pixel row whose distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less. Based on the ratio of the number of extracted pixel rows to the number of the predetermined pixel rows, the road marking detection unit 22 that determines the presence / absence of road marking, and the road marking detection unit 22 determines that there is no road marking. Sometimes, a lane boundary line candidate detected by the lane boundary line detection unit 21 is provided with a determination unit 23 that determines the lane boundary line as a lane boundary line.
With such a configuration, road markings such as pedestrian crossings can be detected together with a simple configuration, and lane boundary lines can be detected with high accuracy.

Further, the lane detection method according to the present embodiment includes a step S90 of detecting a lane boundary line candidate from an edge image generated based on an image obtained by imaging a road, and the edge image is moved to the left and right at the approximate center of the edge image. Step S101 for dividing into blocks, and a step of extracting, from a predetermined pixel row of the edge image, a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less. Detected when S113, steps S116 to S118 for determining the presence or absence of road marking based on the ratio of the number of extracted pixel rows to the number of the predetermined pixel rows, and no road marking This includes steps S121 and S122 for determining a lane boundary candidate as a lane boundary.
With such a configuration, road markings such as pedestrian crossings can be easily detected together and lane boundaries can be detected with high accuracy.

This application claims priority based on Japanese Patent Application No. 2015-108199 filed on May 28, 2015, the entire disclosure of which is incorporated herein.

The lane detection system or the lane detection method according to the embodiment is applied to a vehicle or the like and can collectively detect road markings such as a pedestrian crossing and has industrial applicability.

1 Lane detection system 10 Camera 20 Lane detection unit 21 Lane boundary detection unit 22 Road marking detection unit 23 Determination unit 30 Monitor 40 Speaker

Claims

An edge image generated based on an image of a road is divided into left and right blocks at the approximate center of the edge image,
From the predetermined pixel row of the edge image, extract a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less,
A road marking detection apparatus comprising a road marking detection unit that determines whether or not there is a road marking based on a ratio of the number of extracted pixel rows to the number of the predetermined pixel rows.
An edge image generated based on an image of a road is divided into left and right blocks at the approximate center of the edge image,
From the predetermined pixel row of the edge image, extract a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less,
A failure element detection apparatus comprising a failure element detection unit that determines the presence or absence of a failure element including a preceding vehicle based on a ratio of the number of extracted pixel rows to the number of the predetermined pixel rows.
A lane boundary detection unit that detects lane boundary candidates from an edge image generated based on an image of a road;
The edge image is divided into left and right blocks at approximately the center of the edge image, and a distance between a predetermined pixel row of the edge image and a pixel closest to the approximately center of each of the left and right blocks in the pixel row is A road marking detection unit that extracts pixel rows that are equal to or less than a predetermined value, and determines whether or not there is a road marking based on a ratio of the number of the extracted pixel rows to the number of the predetermined pixel rows;
A lane detection device comprising: a determination unit that determines that the lane boundary candidate detected by the lane boundary detection unit is a lane boundary when the road marking detection unit determines that there is no road marking.
When the lane boundary detection unit detects one lane boundary line candidate and the road marking detection unit determines that there is no road marking,
The lane detection device according to claim 3, wherein the determination unit determines the detected lane boundary line candidate as a lane boundary line and estimates a position of the other lane boundary line.
Detecting a lane boundary line candidate from an edge image generated based on an image of a road;
Dividing the edge image into left and right blocks at substantially the center of the edge image;
Extracting from the predetermined pixel row of the edge image a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less;
Determining the presence or absence of road marking based on the ratio of the number of extracted pixel rows to the number of the predetermined pixel rows;
A lane detection method comprising: determining that the detected lane boundary line candidate is a lane boundary line when it is determined that there is no road marking.
On the computer,
A procedure for detecting lane boundary candidates from an edge image generated based on an image of a road;
Dividing the edge image into left and right blocks at substantially the center of the edge image;
A procedure for extracting, from a predetermined pixel row of the edge image, a pixel row in which the distance between the pixels closest to the approximate center of each of the left and right blocks in the pixel row is a predetermined value or less;
A procedure for determining the presence or absence of a road marking based on the ratio of the number of extracted pixel rows to the number of the predetermined pixel rows;
A program for executing a procedure for determining, when it is determined that there is no road marking, the detected lane boundary line candidate as a lane boundary line.