JP2019020957A - Detection device - Google Patents

Abstract

To provide a technique for appropriately detecting a lane mark defining a traveling road on which a vehicle travels.SOLUTION: A detection unit includes an image acquisition unit S100, a candidate extraction unit S110, a luminance determination unit S160, a plurality of determination units S190, a recognition unit S210, and an output unit S40. The candidate extraction unit extracts at least one line candidate that is a candidate for a lane marking line from a photographed image on the basis of a feature point. The luminance determination unit determines whether or not luminance of a line candidate is equal to or higher than a predetermined luminance threshold value for each of at least one line candidate. When there are a plurality of candidates to be defined which are line candidates and whose luminance is determined to be equal to or higher than the luminance threshold value, a recognition unit recognizes a candidate closest to a vehicle among the plurality of candidates to be defined as a lane marking line.SELECTED DRAWING: Figure 7

Description

  The present disclosure relates to a technique for recognizing a lane mark from an image.

  In Patent Document 1, when a plurality of line candidates that are candidates for lane marks are detected from an image captured by a camera mounted on a vehicle, the line candidate having the highest luminance among the plurality of line candidates is selected as a lane. A technique of recognizing as a mark has been proposed.

Japanese Patent Laying-Open No. 2015-197829

  However, the lane mark may include a line having a lower luminance than the white line, such as a yellow line. As a result of detailed examination by the inventors, the technique described in Patent Document 1 detects a plurality of line candidates having different luminances such as white lines and yellow lines when detecting a lane mark that divides a traveling path on which the vehicle travels. In such a case, a problem has been found that only the white line may be recognized as a lane mark at all times. That is, in the technique described in Patent Document 1, a problem has been found that there is a possibility that a lane mark that divides a travel path on which the vehicle travels may not be detected appropriately.

  One aspect of the present disclosure provides a technique for appropriately detecting lane marks that divide a traveling path on which a vehicle travels.

  One aspect of the present disclosure is a detection device (30) mounted on a vehicle (70). The detection device includes an image acquisition unit (S100), a candidate extraction unit (S110), a luminance determination unit (S160), a multiple determination unit (S190), a recognition unit (S210), an output unit (S40), Is provided.

  The image acquisition unit (S100) is configured to repeatedly acquire captured images that are images captured by at least one camera mounted on a vehicle and that include a travel path that is a road on which the vehicle travels. The candidate extraction unit (S110) is at least one lane line candidate that divides the lane from the captured image based on a feature point that represents a point where the luminance changes between adjacent pixels in the captured image by a predetermined change threshold or more. It is configured to extract one line candidate.

  The luminance determination unit (S160) is configured to acquire the luminance of the line candidate for each of the line candidates and determine whether the luminance of the line candidate is equal to or higher than a predetermined luminance threshold. The multiple determination unit (S190) is configured to determine whether or not there are a plurality of partition candidates that are line candidates and for which the luminance determination unit has determined that the luminance of the line candidate is equal to or higher than the luminance threshold.

  The recognizing unit (S210) is a lane line that divides a lane in which the vehicle travels, with the lane candidate closest to the vehicle among the plurality of lane candidates when it is determined that there are a plurality of lane candidates. It is comprised so that it may recognize as a travel division line. The output unit (S40) outputs a recognition result by the recognition unit.

  According to such a configuration, when a plurality of lane candidates are detected, the lane candidate closest to the vehicle is recognized as a lane line. Can be detected.

  Note that the reference numerals in parentheses described in this column and in the claims indicate the correspondence with the specific means described in the embodiment described later as one aspect, and the technical scope of the present disclosure It is not limited.

The block diagram which shows the structure of a driving assistance system. Explanatory drawing which shows an example of the arrangement position of a camera. Explanatory drawing which shows the function of ECU. The schematic diagram of a bird's-eye view image. Explanatory drawing which shows the brightness | luminance of the division line and repair trace in the area | region A of FIG. The flowchart of a detection process. The flowchart of a recognition process. Explanatory drawing explaining a brightness | luminance threshold value. Explanatory drawing which shows the other example of the arrangement position of a camera.

Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.
[1. Constitution]
[1-1. overall structure]
A driving support system 1 according to this embodiment shown in FIG. 1 is mounted on a vehicle 70. The driving support system 1 recognizes a road marking line from an image taken around the vehicle 70, estimates a road parameter, and executes various types of driving support based on the estimated road parameter. The lane marking is a so-called lane marker, and is a white line or a yellow line drawn on the road surface so as to divide the lane of the traveling road. A travel path is a road on which a vehicle travels. The travel path parameter represents the state of the travel path with respect to the vehicle 70 and the shape of the travel path.

The driving support system 1 includes an ECU 30. The driving support system 1 may include a camera 10, a sensor group 20, and a vehicle control device 50.
The camera 10 includes a front camera 11, a left side camera 12, a right side camera 13, and a rear camera 14. Each of the cameras 11 to 14 can be configured using a CCD image sensor, a CMOS image sensor, or the like.

  As shown in FIG. 2, the front camera 11 is installed, for example, on a bumper at the front end of the vehicle so that the road surface in front of the vehicle is within the shooting range. The left side camera 12 is installed, for example, on the left side mirror so that the road surface on the left side of the vehicle falls within the shooting range. The right side camera 13 is installed, for example, on the right side mirror so that the road surface on the right side of the vehicle is in the shooting range. For example, the rear camera 14 is installed in a bumper at the rear end of the vehicle such that the road surface behind the vehicle is within the shooting range.

Each of the cameras 11 to 14 repeatedly shoots at a preset time interval, for example, every several tens of msec, and outputs a camera image to the ECU 30.
The sensor group 20 is a sensor that measures the behavior of the vehicle 70. Specifically, the sensor group 20 includes a vehicle speed sensor that measures the vehicle speed of the vehicle 70, a yaw rate sensor that measures the yaw rate of the vehicle 70, and the like. The sensor group 20 includes a GPS sensor that detects the position of the vehicle 70, a radar sensor that detects a distance and relative speed with a target existing around the vehicle 70, and an illuminance sensor that detects the brightness around the vehicle 70. Etc. may be provided. The sensor group 20 outputs a detection result such as a sensor value to the ECU 30.

  The vehicle control device 50 includes a microcomputer having a semiconductor memory such as a CPU, a ROM, a RAM, and a flash memory. The vehicle control device 50 controls the steering, braking, engine, and the like of the vehicle 70 so that the vehicle 70 travels in the lane based on the lane marking recognition result output from the ECU 30. The lane marking recognition result may include a travel path parameter.

  The ECU 30 includes a microcomputer having a CPU 41 and a semiconductor memory (hereinafter, memory 42) such as a RAM, a ROM, and a flash memory. Various functions of the ECU 30 are realized by the CPU 41 executing a program stored in a non-transitional physical recording medium. In this example, the memory 42 corresponds to a non-transitional tangible recording medium that stores a program. Also, by executing this program, a method corresponding to the program is executed. Note that the ECU 30 may include one microcomputer or a plurality of microcomputers.

  As shown in FIG. 3, the ECU 30 includes an input processing unit 31, a synthesis processing unit 32, a recognition processing unit 33, and an output processing unit 34 as functions configured by the CPU 41 executing a program. The recognition processing unit 33 includes an output unit image acquisition unit, a candidate extraction unit, a luminance determination unit, a plurality of determination units, a recognition unit, an accumulation unit, an update unit, an environment determination unit, and an output unit. The method of realizing these elements is not limited to software, and some or all of the elements may be realized using hardware that combines a logic circuit, an analog circuit, and the like.

[1-2. Characteristics of the road in the photographed image]
In FIG. 4, the schematic diagram of the bird's-eye view image produced | generated by the synthetic | combination process part 32 is shown. A region indicated by a broken line in the center of the figure is a vehicle region where the vehicle 70 exists. The composition processing unit 32 synthesizes camera images photographed by the four cameras 11-14 and converts them into a bird's eye view, and generates a bird's eye view image surrounding the vehicle area. In the bird's-eye view image shown in FIG. 4, a solid yellow line, a solid white line, and a repair trace exist on the traveling road.

  Here, the solid yellow line and the solid white line are lane markings. Usually, the brightness of the white line is larger than the brightness of the yellow line. Therefore, if the lane marking with the highest luminance is always detected as the traveling lane marking in the captured image, the yellow line may not be detected as the traveling lane marking. The traveling lane line represents a lane line that divides the lane in which the vehicle travels.

  On the other hand, the repair trace is a trace of repairing a crack on the road surface. The repair trace may include, for example, a trace of repairing a crack on an asphalt road surface with tar, or a trace of repairing a crack on a concrete road surface with asphalt. The cracks on the road surface often occur along the lane line due to tire pressure, particularly in snowy areas such as North America, and the repaired trace often becomes a straight line along the lane line.

  Usually, the brightness of the repaired trace is smaller than the brightness of the partition lines such as the white line and the yellow line. In addition, the brightness of the repair trace is smaller than the brightness of the road surface. In general, a straight line obtained by subjecting a feature point extracted from a captured image to processing such as Hough transform is recognized as a division line. The feature point here means an edge point. An edge point refers to a point in the captured image where the brightness changes between adjacent pixels by a change threshold value that is a predetermined threshold value.

  Therefore, as shown in FIGS. 4 and 5, for example, a plurality of repair traces extending substantially parallel to the lane markings are positioned with the road surface therebetween, and the interval between the plurality of repair traces is substantially the same as the width of the lane markings. In some cases, a road surface sandwiched between a plurality of repair traces may be erroneously detected as a lane marking. This is because, if there is a portion having a lower brightness than the road surface, such as a repair mark, this lower brightness portion is erroneously recognized as a road surface, and the original road surface is erroneously recognized as a lane marking.

  The low luminance region referred to below is each of a plurality of linear regions located on the road surface so as to be parallel to the lane markings, and are separated from each other by about the same width as the lane marking lines. Represents a low luminance area. The low-luminance region can include the above-described repair trace. Further, the low luminance region may include cracks on the road surface, tire marks, and the like.

  Note that “parallel” here is not limited to parallel in a strict sense, and may not be strictly parallel as long as there is a possibility of erroneous recognition in the same manner as described above. Also, the line here is not limited to a line in a strict sense, and may not be a line as long as there is a possibility of being erroneously recognized as described above. In addition, the same width as the lane marking here is not limited to the same width as the lane marking in a strict sense. The width may not be the same.

The non-compartment line area referred to below is an area sandwiched between a plurality of low luminance areas on the road surface, and is an area substantially parallel to the division line that may be erroneously recognized as a division line.
The photographed image of the traveling road has such characteristics. Therefore, in the present embodiment, the ECU 30 is configured to appropriately detect the travel lane marking regardless of the luminance difference of the lane marking by executing a detection process described later. Moreover, ECU30 is comprised so that the non-compartment line area | region and lane marking by a repair trace etc. may be discriminate | determined by performing a detection process.

[2. processing]
[2-1. Detection process]
Next, detection processing executed by the ECU 30 will be described with reference to the flowchart of FIG. The detection process represents a processing procedure for recognizing and outputting a travel lane line based on the above-described feature in the captured image of the travel road. This detection process is repeatedly executed at an imaging time interval of the camera 11-14.

  First, in step (hereinafter referred to as S) 10, the ECU 30 acquires a camera image captured by the camera 11-14, and converts the acquired camera image into a digital signal by sampling.

  In S <b> 20, the ECU 30 converts the four digital camera images into a bird's-eye view viewed from a preset virtual viewpoint, and generates a bird's-eye view image showing the surroundings of the vehicle 70.

In S30, the ECU 30 recognizes the travel lane marking from the bird's eye view image generated in step S20 by executing recognition processing. Details of the recognition process will be described later.
In S40, the ECU 30 estimates the travel path parameter of the travel lane line recognized in the recognition process. Then, the ECU 30 outputs the travel path parameter to the vehicle control device 50 via the in-vehicle network as a recognition result of the travel lane marking. The travel path parameters may include, for example, the curvature of the lane line, the lane width, and the angle formed by the traveling direction of the vehicle 70 and the tangential direction of the lane line. ECU30 complete | finishes this recognition process above.

  In the above embodiment, S10 corresponds to the processing by the input processing unit 31, S20 corresponds to the processing by the synthesis processing unit 32, S30 corresponds to the processing by the recognition processing unit 33, and S40 by the output processing unit 34. It corresponds to processing.

[2-2. Recognition process]
Next, the process procedure of the recognition process which ECU30 performs in S30 of a detection process is demonstrated using the flowchart of FIG.

  First, in S100, a captured image is acquired. The photographed image here is an image photographed by at least one camera 10 mounted on the vehicle 70 and including a travel path. Specifically, the ECU 30 acquires the bird's eye view image generated in S20 as a captured image.

  In S110, the ECU 30 extracts line candidates from the captured image based on the edge points in the captured image. A line candidate is a candidate for a lane line that divides a lane. The change threshold for extracting the edge point is stored in the memory 42 in advance.

  Specifically, the ECU 30 detects edge points by applying a sobel filter or the like to the bird's eye view image acquired in S100. Then, the ECU 30 detects a straight line by applying a Hough transform or the like to the detected edge point, and extracts the detected straight line as a line candidate. The ECU 30 extracts line candidates on the left side and the right side of the vehicle 70, respectively. The extracted line candidates may be extracted from the vehicle 70 within a predetermined range. Specifically, line candidates existing outside a predetermined range from the vehicle 70 are highly likely to be adjacent lanes, and therefore, by excluding such line candidates, the processing load can be reduced. .

  In S120, the ECU 30 acquires the brightness of the line candidate. The line candidate luminance represents the luminance value of the line candidate extracted in S110. Specifically, the ECU 30 calculates an average value of luminance values for all pixels included in the line candidate in the bird's eye view image, and acquires the average value as the luminance of the line candidate. However, the method of calculating the brightness of the line candidate is not limited to this. The ECU 30 may be configured to calculate an average value of luminance values for any of a plurality of pixels included in the line candidate and acquire the average value as the luminance of the line candidate. When there are a plurality of line candidates on the left side and the right side of the vehicle 70, the ECU 30 acquires the brightness of the line candidates for each of the plurality of line candidates.

  In S130, the ECU 30 acquires a luminance threshold value. The luminance threshold is a value that is determined in advance and is used to discriminate between lane markings and non-lane marking areas. The luminance threshold value is stored in the memory 42 as will be described later.

  In S140, the ECU 30 determines whether or not the brightness of the traveling road around the vehicle 70 (hereinafter referred to as the vehicle periphery) satisfies a predetermined condition. It should be noted that the brightness of the road around the vehicle referred to here may include the brightness around the vehicle 70 in addition to the brightness of the road around the vehicle 70 itself. Hereinafter, the brightness of the road around the vehicle is simply referred to as the brightness of the road. The ECU 30 shifts the process to S150 when the brightness of the travel path satisfies a predetermined condition, and shifts the process to S160 when the brightness of the travel path does not satisfy the predetermined condition.

  This S140 is a process for determining whether or not to execute S160-S180 and S240-S250 described later. As will be described later, the processing of S160 to S180 is processing for excluding line candidates that are line candidates whose luminance is less than the luminance threshold from the block candidates. The process of S240-S250 is a process for updating the brightness threshold, as will be described later.

  In the present embodiment, the brightness of the traveling road is such that the difference between the brightness of the lane marking and the brightness of the road surface is less than a predetermined threshold (hereinafter, bright threshold) (hereinafter, bright condition). Let it be a predetermined condition. For example, a concrete traveling road in clear weather is an example in which the brightness of the traveling road satisfies the light conditions. In this case, due to the influence of sunlight, the brightness of the road surface may have the same magnitude as the brightness of the dividing lines such as white lines and yellow lines. That is, the difference between the brightness of the road surface and the brightness of the lane markings such as white lines and yellow lines can be less than the bright threshold. The bright threshold here may be a value that is sufficiently smaller than the luminance of the road surface or the luminance of the lane marking, for example, close to zero.

  Furthermore, in this embodiment, the brightness of the traveling road is such that the difference between the brightness of the low brightness area and the brightness of the road surface is less than a predetermined threshold (hereinafter referred to as a dark threshold) (hereinafter referred to as a dark condition). ), The above-mentioned predetermined condition. For example, a traveling road at night, a traveling road in a tunnel, and the like are examples in which the brightness of the traveling road satisfies the dark condition. In this case, the road surface may have a luminance equivalent to the repair mark. That is, the difference between the brightness of the road surface and the brightness of the low brightness area such as the repair mark can be less than the dark threshold. The dark threshold here may be a value that is sufficiently smaller than the brightness of the road surface or the brightness of the repaired track, for example, close to 0. The bright threshold value and the dark threshold value may be the same value. The bright threshold and the dark threshold can be confirmed by experiments or the like.

  Specifically, in this step, the ECU 30 determines that the brightness of the travel path does not satisfy a predetermined condition when the brightness threshold acquired in S130 is a value within a predetermined range. In other words, as shown in FIG. 8, when the acquired luminance threshold is equal to or higher than the predetermined upper threshold, or when the acquired luminance threshold is lower than the predetermined lower threshold smaller than the upper threshold, It is determined that the brightness of the travel path satisfies the predetermined condition.

  As will be described later, the luminance threshold is a value proportional to a value based on the luminance of the travel lane line recognized in the past. That is, the brightness threshold value is large when the brightness of the travel road is sufficiently bright to satisfy the light conditions. Further, the brightness threshold value is small when the brightness of the road is sufficiently dark to satisfy the dark condition.

  Therefore, in this step, the ECU 30 determines whether or not the brightness of the travel path satisfies the predetermined condition based on whether or not the luminance threshold is within a predetermined range. That is, when the luminance threshold is equal to or higher than the upper limit threshold, it is determined that the brightness of the traveling road satisfies the bright condition. That is, it is determined that the brightness of the travel road satisfies the predetermined condition. When the brightness threshold is less than the lower limit threshold, it is determined that the brightness of the traveling road satisfies the dark condition. That is, it is determined that the brightness of the travel road satisfies the predetermined condition. The upper and lower thresholds for representing the predetermined range here are determined in advance by experiments or the like and stored in the memory 42.

  In S150, as an exception to the processing in S160 to S180, the ECU 30 sets all line candidates detected in S110 as section candidates described later. Then, the ECU 30 shifts the process to S190.

  In S160, the ECU 30 obtains the luminance of the line candidate for each line candidate, and determines whether the luminance of the line candidate is equal to or higher than the luminance threshold. Here, the ECU 30 shifts the processing to S170 when the luminance of the line candidate is equal to or higher than the luminance threshold. In S170, the ECU 30 sets the line candidate determined to be equal to or greater than the luminance threshold value in S160 as a partition candidate, and causes the process to proceed to S180. On the other hand, the ECU 30 shifts the process to S180 when the luminance of the line candidate is less than the luminance threshold.

  As will be described later, the luminance threshold is a value proportional to a value based on the luminance of the travel lane line recognized in the past. In addition, the luminance threshold is a value close to the luminance of the traveling lane line recognized in the past, and is smaller than the luminance of the traveling lane line recognized in the past.

  For this reason, as shown in FIG. 8, the fact that the brightness of the line candidate is equal to or higher than the brightness threshold indicates that it is more likely to be a travel lane line. In other words, the section candidate represents something that seems to be a traveling section line. In other words, the fact that the luminance of the line candidate is equal to or higher than the luminance threshold means that it is not a non-compartment line area. The process of S160 is a process for discriminating a non-compartment line area from a plurality of line candidates by comparing the brightness of the line candidate with a brightness threshold.

  In S180, ECU 30 repeats the processing of S160-S180 until the determination of S160 is executed for each of all line candidates detected in S110. ECU30 makes a process transfer to S190, when judgment of S160 is performed about all the line candidates detected in S110.

  In S190, the ECU 30 determines whether or not a plurality of section candidates are set in S150 or S170. The ECU 30 shifts the process to S200 when one section candidate is set, and shifts the process to S210 when a plurality of section candidates are set.

In S200, when one section candidate is set, the ECU 30 determines the section candidate as a travel lane line. Then, the ECU 30 shifts the process to S220.
In S210, when a plurality of candidate sections are set, the ECU 30 determines a candidate section closest to the vehicle 70 among the plurality of candidate sections as a travel lane line. Then, the ECU 30 shifts the process to S220. Note that the ECU 30 determines travel lane markings on the right and left sides of the vehicle 70, respectively. That is, the ECU 30 determines a pair of travel lane markings that are estimated to represent the left and right boundaries of the travel path of the vehicle 70. However, the present invention is not limited to this, and the ECU 30 can be configured to determine the travel lane marking on at least one of the right side and the left side of the vehicle 70.

  In S220, the ECU 30 stores the travel lane line brightness in the memory 42. The travel lane line brightness is the brightness of the travel lane line determined in S200 or S210. Specifically, the brightness of the line candidate determined as the travel lane line among the brightness of the line candidates acquired in S120 is used as the travel lane line brightness.

  In S230, as in S140, the ECU 30 determines whether or not the brightness of the travel path satisfies the predetermined condition. The ECU 30 ends the recognition process when the brightness of the travel path satisfies the predetermined condition, and shifts the process to S240 when the brightness of the travel path does not satisfy the predetermined condition.

  However, unlike S140 in this step, the ECU 30 determines that the brightness of the travel path does not satisfy the predetermined condition when the travel lane line brightness stored in the memory 42 in S220 is within a predetermined value. to decide. Specifically, the ECU 30 determines that the brightness of the travel path satisfies the predetermined condition when the travel lane line brightness is equal to or higher than a predetermined first threshold value. Further, the ECU 30 determines that the brightness of the travel path satisfies the predetermined condition when the travel lane marking luminance is less than a second threshold value that is smaller than the first threshold value.

  The first threshold value and the second threshold value are determined in advance by experiments or the like and are stored in the memory 42. The first threshold value is a value corresponding to the upper limit threshold value, and the second threshold value is a value corresponding to the lower limit threshold value.

  In S240, the ECU 30 calculates the average luminance. The average luminance is an average value for a plurality of past luminances. The past luminance is the travel lane line luminance accumulated in the memory 42 in the past. Here, the average value of a plurality of past luminances accumulated in the memory 42 in the latest past predetermined period is calculated as the average luminance.

  In S250, the ECU 30 calculates a reference value. The reference value is a value obtained by multiplying the average luminance by a predetermined multiple α. The multiple α is a value less than 1, and can be set to a value of 0.7 to 0.8, for example. However, the present invention is not limited to this, and the multiple α can be set as appropriate based on experiments or the like. Then, the ECU 30 stores the calculated reference value in the memory 42 as a new brightness threshold. ECU30 complete | finishes this recognition process above.

[3. effect]
According to the first embodiment described in detail above, the following effects are obtained.
(3a) The ECU 30 repeatedly acquires captured images, and in S110, extracts at least one line candidate from the captured images based on the edge points in the captured images. In S160, for each of at least one line candidate, the ECU 30 determines whether the brightness of the line candidate is equal to or higher than the luminance threshold, and sets the line candidate whose luminance is equal to or higher than the luminance threshold as a partition candidate. In S210, the ECU 30 recognizes a zone candidate closest to the vehicle 70 as a running lane line when there are a plurality of zone candidates. In S40, the ECU 30 outputs a recognition result.

  As a result, when a plurality of section candidates are detected, the section candidate closest to the vehicle 70 is specified as the traveling lane line, so that the traveling lane line is appropriately detected regardless of the luminance difference of the lane lines. be able to.

  In addition, since the line candidate whose luminance is lower than the luminance threshold is not set as the division candidate, it is possible to discriminate the non-division line region from the division line. That is, it is suppressed that a non-division line area is erroneously extracted as a division candidate. As a result, a travel lane line can be appropriately detected from the zone candidates.

  (3b) In S220, the ECU 30 stores the travel lane line luminance in the memory 42 every time a travel lane line is recognized in S210. The ECU 30 calculates an average luminance that is an average value for a plurality of past luminances every time the travel lane line luminance is stored in the memory 42 in S250, calculates a reference value that is smaller than the average luminance, The value is stored again in the memory 42 as a new brightness threshold. In S160, the ECU 30 acquires the brightness threshold value stored in the memory 42 in this way, and makes a determination using the brightness threshold value.

  That is, since the brightness threshold is set based on the brightness of a plurality of lane markings recognized in the past, an appropriate brightness threshold that reflects the changing brightness of the surrounding environment of the vehicle 70 is set. Is done. In addition, since the brightness threshold value is appropriately set, it is possible to prevent a non-division line region due to a repair mark or the like from being erroneously extracted as a division candidate. As a result, the travel lane marking can be detected appropriately.

  (3c) In S140 or S230, the ECU 30 is configured to determine whether or not the brightness of the traveling road around the vehicle satisfies a predetermined condition. As a result, processing according to the brightness of the traveling road around the vehicle is possible.

  (3d) In S250, when it is determined in S250 that the brightness of the travel path satisfies the predetermined condition in S230, the ECU 30 stops storing the reference value in the memory 42 as a new luminance threshold.

  In other words, for example, when the light condition is satisfied and the traveling light is very bright due to strong irradiation light from the sun, or when the dark condition is satisfied and the traveling path is very dark in the tunnel or at night, etc. In this case, the brightness threshold is not updated using the reference value. As described above, the reference value is a value based on the brightness of the travel lane line recognized in the past. The reference value is a large value when the brightness of the travel road is bright, and a small value when the brightness of the travel road is dark.

  Here, if the brightness threshold is updated using the reference value when the travel road is bright enough to satisfy the bright condition, the brightness threshold may be set to a value larger than the normal yellow line brightness. On the other hand, when the brightness threshold is updated using the reference value when the traveling road is dark enough to satisfy the dark condition, the brightness threshold is set to a value smaller than the brightness of the low brightness area such as a normal repair mark. There is a fear. In the former case, the yellow line may not be extracted as a segment candidate. In the latter case, there is a possibility that a low luminance area is extracted as a candidate for a section.

  In the present embodiment, when it is determined that the brightness of the travel path satisfies the predetermined condition, the brightness threshold is not updated using the reference value. As a result, it is possible to prevent the yellow line from being extracted as a segment candidate based on the luminance threshold. Moreover, it can suppress that a low-intensity area | region is extracted as a division candidate based on a brightness | luminance threshold value.

  (3e) In S210, when it is determined in S210 that the brightness of the travel road satisfies the predetermined condition in S140, the ECU 30 is at least one line candidate extracted in S110 and out of all line candidates. The line candidate closest to the vehicle 70 is recognized as a travel lane line.

  That is, when it is determined that the brightness of the travel path satisfies the predetermined condition, the candidate for the segment based on the brightness threshold is not extracted, and the line candidate closest to the vehicle 70 is traveled among all the line candidates. Recognize as a lane marking. As described above, when it is determined that the brightness of the travel path satisfies the predetermined condition, the brightness threshold value is not set appropriately, so that there is a possibility that the block candidates may not be extracted properly. Therefore, in this embodiment, in such a case, the extraction of the partition candidate is stopped. As a result, it is possible to prevent the travel lane line from being erroneously recognized based on the zone candidates.

  (3f) The ECU 30 determines in S140 or S230 that the brightness of the travel road is such that the difference between the brightness of the lane marking and the brightness of the road surface is less than a predetermined light threshold. , Configured to make a decision.

  That is, when the difference between the luminance of the lane line and the luminance of the road surface is less than the bright threshold value, for example, it is possible to stop extracting the candidate candidates or to stop updating the luminance threshold value. In addition, when the difference between the luminance of the lane marking and the luminance of the road surface is less than the bright threshold, it is possible to perform processing for stopping the output of the recognition result. As a result, misrecognition of travel lane markings can be suppressed. Moreover, it can suppress that an incorrect recognition result is output.

  (3g) In S140 or S230, the ECU 30 determines that the brightness of the travel road is such that the difference between the brightness of the low brightness area and the brightness of the road surface is less than a predetermined dark threshold. Is configured to make a determination.

  In other words, when the difference between the luminance of the low luminance region and the luminance of the road surface is less than the dark threshold, for example, it is possible to stop the extraction of the partition candidate or to stop the updating of the luminance threshold. In addition, when the difference between the brightness of the low brightness area and the brightness of the road surface is less than the dark threshold value, it is possible to perform processing to stop outputting the recognition result. As a result, misrecognition of travel lane markings can be suppressed. Moreover, it can suppress that an incorrect recognition result is output.

  (3h) In S140, the ECU 30 determines that the brightness of the travel road satisfies the predetermined condition when the luminance threshold is equal to or greater than the predetermined upper limit threshold. In S140, the ECU 30 determines that the brightness of the travel road satisfies the predetermined condition when the luminance threshold is less than the predetermined lower threshold smaller than the upper threshold. As a result, in order to detect the brightness of the road, it is possible to easily determine whether the brightness of the road satisfies the predetermined condition without providing a new configuration such as an illuminance sensor. Can do.

  (3i) In S230, the ECU 30 determines that the brightness of the travel path satisfies the predetermined condition when the travel lane marking luminance is equal to or higher than a predetermined first threshold value. In S230, the ECU 30 determines that the brightness of the travel path satisfies the predetermined condition when the travel lane line brightness is less than a second threshold value that is smaller than the first threshold value. As a result, the same effect as the above (3h) can be obtained.

  In the above embodiment, the ECU 30 corresponds to the detection device, and the memory 42 corresponds to the storage unit. S40 corresponds to processing by the output unit, S100 corresponds to processing by the image acquisition unit, S110 corresponds to processing by the candidate extraction unit, S140 and S230 correspond to processing by the environment determination unit, and S160 represents luminance. This corresponds to processing by the determination unit. Further, S190 corresponds to processing by a plurality of determination units, S210 corresponds to processing by a recognition unit, S220 corresponds to processing by an accumulation unit, and S250 corresponds to processing by an updating unit.

The bird's-eye view image corresponds to the captured image, and the edge point corresponds to the feature point. Each of the light condition and the dark condition corresponds to a predetermined condition.
[4. Other Embodiments]
As mentioned above, although embodiment of this indication was described, this indication is not limited to the above-mentioned embodiment, and can carry out various modifications.

  (4a) In the above embodiment, based on whether or not the brightness threshold is within a predetermined range in S140, and based on whether or not the travel division linearity is within a predetermined range in S230, It has been determined whether the brightness satisfies the predetermined condition. However, it is not limited to this.

  For example, in the vehicle 70, an illuminance sensor that detects ambient brightness may be mounted as a sensor of the sensor group 20. The ECU 30 may be configured to determine that the brightness of the travel path satisfies the predetermined condition when the sensor value of the illuminance sensor is equal to or greater than a predetermined first illuminance threshold. Further, when the sensor value of the illuminance sensor is less than the second illuminance threshold value that is smaller than the first illuminance threshold value, the ECU 30 determines that the brightness of the travel path satisfies the predetermined condition. It may be configured to determine.

  Further, for example, in the vehicle 70, a GPS device may be mounted as a sensor of the sensor group 20, or a map database that stores map data may be mounted. When the ECU 30 determines that the current position of the vehicle 70 is in the tunnel based on the detection result of the GPS device and the map data, the brightness of the travel path satisfies the predetermined condition. It may be configured to determine.

  (4b) As described above, when it is determined that the brightness of the travel path satisfies the predetermined condition, the brightness threshold value is not appropriately set, and the travel lane line may not be properly determined. Thus, in S40, the ECU 30 may be configured to stop outputting the recognition result when it is determined in S140 or S230 that the brightness of the travel path satisfies the predetermined condition. As a result, an erroneous recognition result can be prevented from being output. In addition, it is possible to suppress erroneous vehicle control based on an incorrect travel path parameter.

  (4c) In the above embodiment, the ECU 30 is configured to recognize the travel lane marking from the bird's-eye view image using the bird's-eye view image generated based on the camera image of the camera 11-14 as a photographed image. The image is not limited to a bird's eye view image. For example, as shown in FIG. 9, one camera 10 may be mounted on the vehicle so as to photograph a road surface ahead of the vehicle 70. Specifically, the camera 10 is attached to the center front side of the vehicle 70, and is configured to photograph a region extending in a predetermined angle range toward the front of the vehicle 70 as shown in FIG. Also good. The ECU 30 may be configured to detect an edge point from a photographed image photographed by the camera 10 and recognize a travel lane line from the photographed image based on the edge point, as in the above embodiment. good.

  (4d) A plurality of functions of one constituent element in the above embodiment may be realized by a plurality of constituent elements, or a single function of one constituent element may be realized by a plurality of constituent elements. . Further, a plurality of functions possessed by a plurality of constituent elements may be realized by one constituent element, or one function realized by a plurality of constituent elements may be realized by one constituent element. Moreover, you may abbreviate | omit a part of structure of the said embodiment. In addition, at least a part of the configuration of the above embodiment may be added to or replaced with the configuration of the other embodiment. In addition, all the aspects included in the technical idea specified from the wording described in the claims are embodiments of the present disclosure.

  (4e) In addition to the ECU 30 and the driving support system 1 described above, the present invention can be implemented in various forms such as a program for causing the ECU 30 to function, a non-transition actual recording medium such as a semiconductor memory storing the program, a lane marking detection method Disclosure can also be realized.

  10 cameras, 30 ECUs, 42 memories, 70 vehicles.

Claims (7)

A detection device (30) mounted on a vehicle (70),
An image acquisition unit (S100) configured to repeatedly acquire an image captured by at least one camera (10) mounted on the vehicle and including a travel path that is a road on which the vehicle travels. )When,
At least one line candidate that is a candidate for a lane marking that divides a lane is extracted from the photographed image based on a feature point that represents a point where the luminance changes between adjacent pixels in the photographed image by a predetermined change threshold or more. A candidate extraction unit (S110) configured as described above,
For each of the line candidates, a luminance determining unit (S160) configured to acquire the luminance of the line candidate and determine whether the luminance of the line candidate is equal to or higher than a predetermined luminance threshold;
A plurality of determination units (S190) configured to determine whether or not there are a plurality of partition candidates that are the line candidates and the luminance determination unit has determined that the luminance of the line candidates is equal to or higher than the luminance threshold;
When the plurality of determination units determine that there are a plurality of division candidates, the division candidate that is the division line that divides the lane in which the vehicle runs is selected as the division candidate closest to the vehicle among the plurality of division candidates. A recognition unit (S210) configured to recognize as a line;
An output unit (S40) for outputting a recognition result by the recognition unit;
A detection device comprising: The detection device according to claim 1,
An accumulating unit (S220) configured to store in the storage unit (42) the running lane line luminance representing the luminance of the traveling lane line recognized by the recognition unit;
An average brightness that is an average value for a plurality of past brightnesses that are the travel lane line brightness that has been stored in the storage unit in the past is acquired, and is a value that is smaller than the average brightness based on the average brightness An update unit (S250) that calculates a reference value and re-accumulates the reference value in the storage unit as a new luminance threshold; the luminance determination unit acquires the luminance threshold from the storage unit; A detection device configured to determine. The detection device according to claim 2,
An environment determination unit (S140, S230) configured to determine whether the brightness of the road around the vehicle satisfies a predetermined condition;
A detection device further comprising: The detection device according to claim 3,
The update unit stores the reference value as the new brightness threshold in the storage unit when the environment determination unit (S230) determines that the brightness of the road around the vehicle satisfies the predetermined condition. A detection device configured to stop memorizing. The detection device according to any one of claims 3 and 4,
The recognizing unit, when the environment determining unit (S140) determines that the brightness of the road around the vehicle satisfies the predetermined condition, all the at least one extracted by the candidate extracting unit A detection device configured to recognize a line candidate closest to the vehicle among the line candidates as the travel lane marking. The detection device according to any one of claims 3 to 5,
The environment determination unit, as the predetermined condition, that the brightness of the road around the vehicle is such that the difference between the brightness of the lane marking and the brightness of the road surface is less than a predetermined brightness threshold, A detection device configured to determine. The detection device according to any one of claims 3 to 6,
The environment determination unit, as the predetermined condition, that the brightness of the road around the vehicle is such that the difference between the luminance of the low luminance area and the luminance of the road surface is less than a predetermined dark threshold value. Configured to judge,
The low luminance region is a plurality of linear regions located on the road surface so as to be parallel to the partition line, and represents a region having a lower luminance than the road surface. apparatus.

Images