CN107918763A - Method for detecting lane lines and system - Google Patents
Method for detecting lane lines and system Download PDFInfo
- Publication number
- CN107918763A CN107918763A CN201711070348.8A CN201711070348A CN107918763A CN 107918763 A CN107918763 A CN 107918763A CN 201711070348 A CN201711070348 A CN 201711070348A CN 107918763 A CN107918763 A CN 107918763A
- Authority
- CN
- China
- Prior art keywords
- line
- lane
- lane line
- detection
- angle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000001514 detection method Methods 0.000 claims abstract description 91
- 238000001914 filtration Methods 0.000 claims abstract description 11
- 238000004590 computer program Methods 0.000 claims description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000002262 irrigation Effects 0.000 description 2
- 238000003973 irrigation Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of method for detecting lane lines, including：Gather road image；Line segment detection is carried out to the road image；The line segment detected is filtered, horizontal disturbance line is removed；The horizontal disturbance line is the line segment of angle within a preset range between horizontal line；End point is determined according to the line segment after filtering；Calculate registration of the scan line under each scanning angle with all line segments Jing Guo the end point；Angle of the scanning angle between scan line and horizontal line；Registration peak value is determined according to the registration under each scanning angle；And by the straight line segments recognition corresponding to the corresponding scanning angle of each registration peak value it is lane line.Above-mentioned method for detecting lane lines determines that the process of registration peak value can be detected multiple lane lines by the registration under each scanning angle, and is not limited to adjacent two track, so as to fulfill Multi-lane Lines Detection.
Description
Technical Field
The invention relates to the technical field of intelligent driving, in particular to a lane line detection method and system.
Background
With the development of economy and science and technology in recent years, the living standard of people is increasingly improved, the traffic facilities of cities are continuously improved, and the number of automobiles is obviously increased. Meanwhile, traffic safety accidents caused by automobiles are increasing. The development of an efficient, safe and comfortable unmanned intelligent driving automobile is a feasible scheme for reducing traffic safety accidents. An unmanned intelligent driving automobile is a complex intelligent control system and comprises a plurality of modules such as mechanical control, path planning, path decision and environment perception. The lane line detection is an important part in the environment sensing module, and can be used for early warning of vehicle driving deviation and providing decision information for lane changing. The traditional lane line detection method can only detect two lanes nearest to the automobile, but cannot realize multi-lane line detection, so that the actual driving requirement cannot be met.
Disclosure of Invention
In view of the above, it is necessary to provide a lane line detection method and system capable of realizing multi-lane line detection.
A lane line detection method includes:
collecting a road image;
carrying out line segment detection on the road image;
filtering the detected line segments to remove horizontal interference lines; the horizontal interference line is a line segment, and an included angle between the horizontal interference line and the horizontal line is within a preset range;
determining a vanishing point according to the filtered line segment;
calculating the contact ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle; the scanning angle is an included angle between a scanning line and a horizontal line;
determining a contact ratio peak value according to the contact ratio under each scanning angle; and
and identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as a lane line.
According to the lane line detection method, the line segment detection is carried out on the collected road image, the horizontal interference line is removed, and then the vanishing point is determined according to the filtered line segment. After the vanishing point is determined, the contact ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle is calculated, so that a contact ratio peak value is determined according to each contact ratio, and the line segment corresponding to the scanning angle corresponding to each contact ratio peak value is identified as the lane line. The process of determining the peak value of the contact ratio through the contact ratio at each scanning angle can detect a plurality of lane lines without being limited to two adjacent lanes, so that the detection of the plurality of lane lines is realized.
In one embodiment, the step of determining the vanishing points according to the filtered line segments comprises:
solving the distribution probability of the intersection points of all the line segment pairs on the image plane; and
and determining the intersection point with the highest distribution probability on the image plane as a vanishing point.
In one embodiment, after the step of determining a vanishing point according to the filtered line segments and before the step of calculating the overlap ratio of the scan line passing through the vanishing point and all the line segments at each scan angle, the method further includes: and filtering the line segment passing through the vanishing point and the line segment which does not extend towards the vanishing point.
In one embodiment, the step of identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak as the lane line includes:
identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as an initial lane line; and
and taking the lane line obtained after the denoising processing is carried out on the initial lane line as a final lane line detection and identification result.
In one embodiment, after the step of identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak as the initial lane line and before the step of performing denoising processing on the initial lane line as the final lane line detection and identification result, the method further includes: and adjusting the angle of each initial lane line based on the horizontal line formed by the vanishing point.
In one embodiment, the step of using the lane line obtained by denoising the initial lane line as the final lane line detection and identification result includes:
determining a lane line scheme when the included angles between two adjacent lane lines are both larger than an angle threshold according to the initial lane line, and determining the minimum adjacent lane line included angle in the corresponding lane line scheme; and
and taking the result of the lane line scheme with the largest minimum adjacent lane line included angle as a final lane line detection and identification result.
In one embodiment, after the step of determining a vanishing point according to the filtered line segments and before the step of calculating the overlap ratio of the scan line passing through the vanishing point and all the line segments at each scan angle, the method further includes:
determining the rotation angle of the vehicle by combining the previously collected road image; and
and adjusting the vanishing point according to the rotation angle.
A lane line detection method includes:
collecting a road image;
carrying out line segment detection on the road image;
filtering the detected line segments to remove horizontal interference lines; the horizontal interference line is a line segment, and an included angle between the horizontal interference line and the horizontal line is within a preset range;
determining a vanishing point according to the filtered line segment;
calculating the contact ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle; the scanning angle is an included angle between a scanning line and a horizontal line;
determining a contact ratio peak value according to the contact ratio under each scanning angle;
identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as an initial lane line;
matching each initial lane line with the lane lines in the lane line set; the lane lines of the lane line set consist of initial lane lines detected by the road images acquired every time before;
counting the matching times of the successfully matched lane lines in the lane line set as the detection times; and
taking the lane line with the detection times larger than the detection threshold value in the lane line set as a corrected initial lane line; and
and denoising the corrected initial lane line to obtain a final lane line detection and identification result.
In one embodiment, the step of counting the matching times of the lane lines successfully matched in the lane line set, and the step of detecting the number of times further includes:
zeroing the lost times of the successfully matched lane lines in the lane line set; adding one to the number of times of losing the unsuccessfully matched lane line in the lane line set; setting the detection times of the lane lines as one and setting the loss times as zero after the initial lane lines which are identified in the step of identifying the line segments corresponding to the scanning angles corresponding to the peak values of the contact ratio as the initial lane lines and do not appear in the lane line set are added to the lane line set as new elements;
and when the number of times of losing the lane lines in the lane line set is greater than a loss threshold value, removing the lane lines from the lane line set.
A lane line detection system includes a memory and a processor; the memory has stored therein a computer program; wherein the computer program, when executed by the processor, causes the processor to perform the steps of the method according to any of the preceding embodiments.
Drawings
FIG. 1 is a flow chart of a lane line detection method in one embodiment;
FIG. 2 is a schematic diagram illustrating an embodiment of a scan line during rotational scanning;
FIG. 3 is a graph illustrating the relationship between the overlap ratio and the scan line at each scan angle according to an embodiment;
FIG. 4 is a schematic diagram of a detection result obtained by the lane marking detection method shown in FIG. 1;
FIG. 5 is a flowchart of step S140 in one embodiment;
fig. 6 is a lane line detection result in which the non-vanishing point is adjusted;
FIG. 7 is a lane line detection result after adjusting the vanishing points;
FIG. 8 is a flowchart of step S170 in one embodiment;
FIG. 9 is a diagram illustrating the existence of noise segments in an initial lane line determined based on a peak overlap ratio in one embodiment;
FIG. 10 is a schematic diagram illustrating adjustment of an angle of an initial lane line in one embodiment;
FIG. 11 is a flowchart of step S330 in one embodiment;
fig. 12 is a flowchart of a lane line detection method in another embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The lane line detection method in one embodiment can realize the detection of multiple lane lines and can be applied to the technical field of automatic and auxiliary driving. Fig. 1 is a flowchart of a lane line detection method in an embodiment, the method including the steps of:
step S110, collecting road images.
The method comprises the following steps that a front road video can be shot through a camera device on the vehicle in the driving process of the vehicle, so that a road image corresponding to a current frame is collected according to the shot video.
Step S120, a line segment is detected for the road image.
In one embodiment, the road image collected by the camera device is mostly a color image, such as an RGB image. Therefore, the RGB image is converted into a gray image before the line segment detection is performed on the road image. Usually, the road surface and the lane line have obvious contrast in gray scale, so that the detection of the lane line is more favorably realized after the road surface and the lane line are converted into gray scale images. In one embodiment, when the image captured directly by the camera device is a gray image, the gray image is directly used without conversion.
In the present embodiment, the detection of a Line Segment in a road image is realized by an LSD (Line Segment Detector, straight Line Segment detection algorithm). When the road is in a curved section, as long as the curvature of the lane line is not particularly large, the lane line can be approximated to a straight line section, so that the detection of each line section can be realized through the algorithm.
Step S130, filtering the detected line segments to remove the horizontal interference lines.
All line segments in the road image are detected in the line segment detection process, and some horizontal line segments or line segments close to the horizontal line segments need to be filtered at the moment, because the line segments cannot be lane lines, the line segments can be prevented from interfering the subsequent detection and identification process, and the time for calculating vanishing points (disappearing points) later is shortened. In this embodiment, the horizontal interference line refers to a line segment having an included angle with the horizontal line within a preset range. The preset range can be set according to actual needs, and it is ensured that no lane line appears in the preset range.
And step S140, determining a vanishing point according to the filtered line segment.
The vanishing point is a point determined by an included angle between the lane lines.
And step S150, calculating the coincidence ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle.
Specifically, a straight line passing through the vanishing point can be customized as the scanning line Lt(theta). The scanning line Lt(θ) performing a rotational scan at a counterclockwise or clockwise angle about the vanishing point. The scanning angle theta is the scanning line Lt(θ) angle to the horizontal, as shown in FIG. 2.
Calculating the scanning line Lt(θ) degree of coincidence with all line segments at each scan angle θ, which may be given by a score SLt(theta) is shown. For example, in one embodiment, the scan line L may be calculated by the following formulat(θ) score at each scan angle θ:
wherein,liindicating the length of the ith line segment; w is aiRepresenting the width of the ith line segment; diRepresents the scanning line Lt(θ) minimum distance from ith line segment;represents the scanning line Lt(theta) the angle between the ith line segment.
The coincidence degree of the scanning line and all line segments under each scanning angle can be calculated through the formula, and the relation graph of each coincidence degree and the scanning angle is shown in fig. 3. It is understood that in other embodiments, other calculation methods may be used to calculate the coincidence degree of the scan line with all line segments at each scan angle.
Step S160, determining a contact ratio peak value according to the contact ratio at each scanning angle.
Since the coincidence degree of the scanning line with a certain line segment of all the line segments is higher, the coincidence degree of the scanning line obtained finally is also higher. During the scanning process of the scanning line, the coincidence degree is higher only when the scanning line at the scanning angle is coincident with or close to coincident with the detected line segment, and the coincidence degree is lower at other scanning angles. Therefore, the peak value of the contact ratio can be determined according to the change situation of the contact ratio. In an embodiment, the coincidence degree corresponding to the peak in the relationship graph may also be used as the peak value of the coincidence degree according to the finally obtained relationship graph of the coincidence degree and the scanning angle. As shown in fig. 3, the position of the black dot is a peak, and the corresponding contact ratio is taken as the peak value of the contact ratio. In another embodiment, the flooding watershed algorithm may also be used to find the overlap ratio peak. Specifically, some points approximate to the wave troughs are found in the oscillogram (for example, there may be several continuous points in the wave troughs, and one point is taken at any time), irrigation is performed from these points to the left and right, the intersection points of irrigation by water in different directions are wave peaks, and the contact ratio corresponding to the wave peaks is the contact ratio peak value.
In step S170, the line segment corresponding to the scanning angle corresponding to each contact ratio peak is identified as the lane line.
And finding out the scanning angle corresponding to the peak value according to the determined contact ratio peak value, thereby identifying the line segment corresponding to the scanning angle as the lane line. The line segment corresponding to the scanning angle is the line segment which has the included angle with the horizontal line and is closest to the scanning angle, or the line segment which has the included angle with the horizontal line and is closest to the scanning line under the scanning angle. Due to the adoption of the method for detection, a plurality of coincidence degree peak values can be detected, and the detection of multiple lane lines is further realized, as shown in fig. 4. The thick lines in fig. 4 are detected line segments, and the thin dotted lines indicate real lane lines captured in the image.
According to the lane line detection method, the line segment detection is carried out on the collected road image, the horizontal interference line is removed, and then the vanishing point is determined according to the filtered line segment. After the vanishing point is determined, the contact ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle is calculated, so that a contact ratio peak value is determined according to each contact ratio, and the line segment corresponding to the scanning angle corresponding to each contact ratio peak value is identified as the lane line. The detection of a plurality of lane lines can be realized through the process of determining the peak value of the contact ratio through the contact ratio at each scanning angle, and the detection is not limited to two adjacent lanes, so that the detection of the plurality of lane lines is realized.
In an embodiment, step S140 may be implemented by the process described in fig. 5, and specifically includes the following steps:
step S210, obtaining the distribution probability of the intersection points of all the line segment pairs on the image plane.
In this embodiment, the distribution probability of the intersection point of a line segment pair consisting of any two line segments among the filtered line segments on the image plane is obtained. In one embodiment, for any pair of line segments, their intersection is considered to be a Gaussian distribution on an image plane. This gaussian distributed covariance matrix is considered with the length and height of the line segment pair as the standard deviation. The distribution probability is also considered as a voting weight for the vanishing points, i.e. each distribution on the image plane is given a weight value.
Step S220, determining the intersection point with the highest distribution probability on the image plane as the vanishing point.
And determining the intersection point with the highest distribution probability as a vanishing point according to the distribution probability of the intersection points of the line segment pairs on the image plane. In one embodiment, when each distribution is given a corresponding weight, the superimposition weight of each point on the image plane may be counted, and the point having the largest superimposition weight is determined as the vanishing point.
In one embodiment, after step S140 and before step S150, a step of determining a rotation angle of the vehicle in combination with the previously collected road image and adjusting the determined vanishing point according to the rotation angle to improve the detection effect of the lane line may be performed, the adjusting process may be performed in a kalman filter manner, so as to improve the detection effect, specifically, for the vanishing point, a state variable is two-dimensional x ═ x, y, where x and y coordinate axes are defined, a control variable is u ═ x, y, the control variable corresponds to the rotation angle of the vehicle, i.e., the rotation angle of the steering wheel, in the present embodiment, a control vector u ═ β ═ gamma, where gamma represents the rotation angle of the steering wheel, β is an adjustable parameter, the rotation angle of the steering wheel or the rotation angle of the vehicle may be determined according to the previously collected road image, i.e., a preset frame image may be obtained from the forward direction, so as to determine the rotation angle of the steering wheel according to the preset frame image, further, the rotation angle of the steering wheel or the rotation angle of the vehicle may be determined according to correct the vanishing point, and the detection result is a thick line, and the detection result is performed for the lost lane line is 7.
In an embodiment, after step S140 and before step S150, a step of filtering the line segment passing through the vanishing point and the line segment that does not extend toward the vanishing point needs to be performed. Since the line segments passing through the vanishing point and the line segments not extending toward the vanishing point are not lane lines, they need to be filtered to ensure that they do not affect the identification process of the following lane lines, such as step S150.
In an embodiment, step S170 may include the steps as shown in fig. 8:
in step S310, the line segment corresponding to the scanning angle corresponding to each contact ratio peak is identified as the initial lane line.
Noise segments (i.e., non-lane segments) may be present in each segment determined according to the overlap ratio peak, as shown in fig. 9. Therefore, when there are some noise line segments, the result cannot be directly regarded as the final detection result, but only the line segments are recognized as the initial lane lines.
And step S330, using the lane line obtained after the denoising processing is carried out on the initial lane line as a final lane line detection and identification result.
And filtering the detected noise line segment in the initial lane line through denoising processing, thereby obtaining a final lane line detection and identification result.
In one embodiment, after step S310 is executed and before step S320 is executed, step S320 is also executed, as shown in fig. 8. Step S320, adjusting the angle of each initial lane line based on the horizontal line formed by the vanishing point. Due to the visual angle error, the included angle formed by two adjacent lane lines at two sides is smaller than that formed by two adjacent lane lines in the middle. In order to make the angle formed by any two adjacent lane lines almost equal, the angle of each initial lane line needs to be adjusted based on the horizontal line formed by the vanishing point. Specifically, the 180-degree sector may be divided into 7 sections according to a horizontal line formed by vanishing points. The angle ranges of the divided parts are respectively 0-25, 25-45, 45-70, 70-110, 110-135, 135-155 and 155-180, and the value of each part is from {1, 2, 4, 6}, as shown in fig. 10. Therefore, it is assumed that there is a lane line with an angle of 35 degrees (in this embodiment, the angle of the lane line refers to the angle between the lane line and the horizontal line). As can be seen from fig. 10, 35 degrees is in the interval of 25 to 45 degrees, and thus the converted angle is (25 × 6+ (35-25) × 4), i.e. 190 degrees.
In an embodiment, step S330 may be implemented by the flow shown in fig. 11, and specifically includes the following steps:
step S332, determining a lane line scheme when the included angles between two adjacent lane lines are both larger than an angle threshold according to the initial lane line, and determining the minimum adjacent lane line included angle in the corresponding lane line scheme.
Since the detected initial lane line may include a noise line segment, an included angle between two adjacent lane lines in the lane line after the angle adjustment should be greater than an angle threshold. The angle threshold is the minimum included angle between two adjacent lane lines, and when the angle threshold is smaller than the angle threshold, at least one of the two adjacent line segments is necessarily determined to be a noise line segment. Therefore, at least one lane line plan satisfying the threshold condition can be determined according to the angle threshold. And when only one lane line scheme meeting the conditions is finally obtained, directly taking the lane line scheme as a final lane line detection and identification result. When it is detected that a plurality of lane lines all satisfy the threshold condition, the minimum value of the included angle between the adjacent lane lines needs to be maximized. In this embodiment, when the lane line scheme meeting the condition is determined, the minimum adjacent lane line included angle in the lane line scheme, that is, the minimum value of the included angle between every two adjacent lane lines, is determined.
In step S334, the lane line plan with the largest minimum adjacent lane line angle is used as the final lane line detection and identification result.
The lane line scheme with the largest and smallest adjacent lane line angle has the lowest possibility of detecting the noise line segment, and therefore the lane line scheme is used as a final detection and identification result.
In one embodiment, step S330 may be implemented by the following pseudo code:
n is the number of lane lines
The ith lane line is adjusted to form an angle with the horizontal line
dpn [ i ] how many lane lines up to the ith lane satisfy the condition
dpv [ i ] how much the minimum angle to the ith case is
The method adopts a dynamic programming method to find an optimal scheme to reserve the most lane lines, thereby realizing the detection process of the multiple lane lines.
Fig. 12 is a flowchart of a lane line detection method in another embodiment, which includes the steps of:
step S402, collecting road images.
In step S404, the road image is subjected to line segment detection.
Step S406, filtering the detected line segments to remove the horizontal interference lines.
And step S408, determining a vanishing point according to the filtered line segment.
Step S410, calculating the coincidence ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle.
In step S412, a contact ratio peak is determined according to the contact ratio at each scanning angle.
In step S414, the line segment corresponding to the scanning angle corresponding to each contact ratio peak is identified as the initial lane line.
In step S416, each initial lane line is matched with a lane line in the lane line set.
In the process of detecting the lane lines by adopting the method, the initial lane lines detected by the acquired road images each time are stored to form a lane line set, namely the lane lines in the lane line set consist of the initial lane lines detected by the acquired road images each time.
And matching each initial lane line with a lane line in the lane line set, namely judging whether each initial lane line is overlapped or nearly overlapped with a certain lane line in the lane line set, and if one lane line in the lane line set is overlapped or nearly overlapped with the lane line, considering that the lane line in the lane line set is matched with the initial lane line.
In step S418, the matching times of the lane lines successfully matched in the lane line set are counted as the detection times.
When it is determined that a certain lane line in the lane line set has an initial lane line matched with the certain lane line, it can be determined that the lane line is successfully matched, so that the matching times of the lane line are counted and used as the detection times. Specifically, the number of times of detection of the lane line in the lane line set is increased by one to count the number of times of occurrence of continuous detection of the lane line.
In step S420, the lane line with the detection frequency greater than the detection threshold in the lane line set is used as the corrected initial lane line.
When the number of times of detection of a lane line in the lane line set is greater than the detection threshold, that is, the number of times of continuous occurrence of the lane line is greater than the detection threshold, the lane line may be used as one of the initial lane lines to perform subsequent denoising processing. The initial lane line is determined by judging the occurrence frequency of the lane line in the continuous frames, so that the accuracy of lane line detection can be improved.
In an embodiment, in step S420, it is further required to set the number of missing times of the successfully matched lane line in the lane line set to zero, and add one to the number of missing times of the unsuccessfully matched lane line in the lane line set. Furthermore, it is also necessary to set the number of times of detection of the lane line to one and zero after the initial lane line identified in step S414 and not appearing in the lane line set is added as a new element to the lane line set. And after the operation is finished, removing the lane lines with the loss times larger than the loss threshold value from the lane line set. By the method, the lane line queue is maintained, the accuracy of each lane line in the lane line set can be ensured, and the accuracy of lane line detection is improved.
Step S422, the corrected initial lane line is denoised to obtain the final lane line detection and identification result.
In this embodiment, the steps in the previous embodiments are not described repeatedly. The method combines the previous frames to adjust the lane line, so that the accurate detection and identification of the lane line can be realized.
An embodiment of the present invention further provides a lane line detection system, which includes a memory and a processor. Wherein the memory has stored therein a computer program which, when executed by the processor, causes the processor to perform the steps of the method as described in any of the previous embodiments.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A lane line detection method includes:
collecting a road image;
carrying out line segment detection on the road image;
filtering the detected line segments to remove horizontal interference lines; the horizontal interference line is a line segment, and an included angle between the horizontal interference line and the horizontal line is within a preset range;
determining a vanishing point according to the filtered line segment;
calculating the contact ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle; the scanning angle is an included angle between a scanning line and a horizontal line;
determining a contact ratio peak value according to the contact ratio under each scanning angle; and
and identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as a lane line.
2. The method of claim 1, wherein the step of determining vanishing points based on the filtered line segments comprises:
solving the distribution probability of the intersection points of all the line segment pairs on the image plane; and
and determining the intersection point with the highest distribution probability on the image plane as a vanishing point.
3. The method according to claim 1, wherein after the step of determining the vanishing point according to the filtered line segments and before the step of calculating the overlap ratio of the scanning line passing through the vanishing point with all the line segments at each scanning angle, the method further comprises: and filtering the line segment passing through the vanishing point and the line segment which does not extend towards the vanishing point.
4. The method according to claim 1, wherein the step of identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as the lane line comprises:
identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as an initial lane line; and
and taking the lane line obtained after the denoising processing is carried out on the initial lane line as a final lane line detection and identification result.
5. The method according to claim 4, wherein after the step of identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as the initial lane line and before the step of denoising the initial lane line as the final lane line detection and identification result, the method further comprises: and adjusting the angle of each initial lane line based on the horizontal line formed by the vanishing point.
6. The method according to claim 4, wherein the step of using the lane line obtained by denoising the initial lane line as a final lane line detection recognition result comprises:
determining a lane line scheme when the included angles between two adjacent lane lines are both larger than an angle threshold according to the initial lane line, and determining the minimum adjacent lane line included angle in the corresponding lane line scheme; and
and taking the result of the lane line scheme with the largest minimum adjacent lane line included angle as a final lane line detection and identification result.
7. The method according to claim 1, wherein after the step of determining the vanishing point according to the filtered line segments and before the step of calculating the overlap ratio of the scanning line passing through the vanishing point with all the line segments at each scanning angle, the method further comprises:
determining the rotation angle of the vehicle by combining the previously collected road image; and
and adjusting the vanishing point according to the rotation angle.
8. A lane line detection method includes:
collecting a road image;
carrying out line segment detection on the road image;
filtering the detected line segments to remove horizontal interference lines; the horizontal interference line is a line segment, and an included angle between the horizontal interference line and the horizontal line is within a preset range;
determining a vanishing point according to the filtered line segment;
calculating the contact ratio of the scanning line passing through the vanishing point and all line segments under each scanning angle; the scanning angle is an included angle between a scanning line and a horizontal line;
determining a contact ratio peak value according to the contact ratio under each scanning angle;
identifying the line segment corresponding to the scanning angle corresponding to each contact ratio peak value as an initial lane line;
matching each initial lane line with the lane lines in the lane line set; the lane lines of the lane line set consist of initial lane lines detected by the road images acquired every time before;
counting the matching times of the successfully matched lane lines in the lane line set as the detection times; and
taking the lane line with the detection times larger than the detection threshold value in the lane line set as a corrected initial lane line; and
and denoising the corrected initial lane line to obtain a final lane line detection and identification result.
9. The method according to claim 8, wherein the step of counting the number of matching times of the lane line successfully matched in the lane line set further comprises:
zeroing the lost times of the successfully matched lane lines in the lane line set; adding one to the number of times of losing the unsuccessfully matched lane line in the lane line set; setting the detection times of the lane lines as one and setting the loss times as zero after the initial lane lines which are identified in the step of identifying the line segments corresponding to the scanning angles corresponding to the peak values of the contact ratio as the initial lane lines and do not appear in the lane line set are added to the lane line set as new elements;
and when the number of times of losing the lane lines in the lane line set is greater than a loss threshold value, removing the lane lines from the lane line set.
10. A lane line detection system includes a memory and a processor; the memory has stored therein a computer program; characterized in that the computer program, when executed by the processor, causes the processor to carry out the steps of the method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711070348.8A CN107918763A (en) | 2017-11-03 | 2017-11-03 | Method for detecting lane lines and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711070348.8A CN107918763A (en) | 2017-11-03 | 2017-11-03 | Method for detecting lane lines and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107918763A true CN107918763A (en) | 2018-04-17 |
Family
ID=61895240
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711070348.8A Pending CN107918763A (en) | 2017-11-03 | 2017-11-03 | Method for detecting lane lines and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107918763A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110838233A (en) * | 2019-10-12 | 2020-02-25 | 中国平安人寿保险股份有限公司 | Vehicle behavior analysis method and device and computer readable storage medium |
| CN113822218A (en) * | 2021-09-30 | 2021-12-21 | 厦门汇利伟业科技有限公司 | Lane line detection method and computer-readable storage medium |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004246641A (en) * | 2003-02-14 | 2004-09-02 | Nissan Motor Co Ltd | Traffic white lane line recognizing device |
| CN101608924A (en) * | 2009-05-20 | 2009-12-23 | 电子科技大学 | A Lane Line Detection Method Based on Gray Level Estimation and Cascaded Hough Transform |
| CN102303609A (en) * | 2011-06-16 | 2012-01-04 | 广东铁将军防盗设备有限公司 | Lane departure warning system and method |
| CN102324017A (en) * | 2011-06-09 | 2012-01-18 | 中国人民解放军国防科学技术大学 | A Lane Line Detection Method Based on FPGA |
| CN102360499A (en) * | 2011-06-30 | 2012-02-22 | 电子科技大学 | Multi-lane line tracking method based on Kalman filter bank |
| US20130202155A1 (en) * | 2012-02-03 | 2013-08-08 | Gopal Gudhur Karanam | Low-cost lane marker detection |
| CN103295420A (en) * | 2013-01-30 | 2013-09-11 | 吉林大学 | Method for recognizing lane line |
| CN103971521A (en) * | 2014-05-19 | 2014-08-06 | 清华大学 | Method and device for detecting road traffic abnormal events in real time |
| CN104376297A (en) * | 2013-08-12 | 2015-02-25 | 株式会社理光 | Detection method and device for linear indication signs on road |
| CN104742912A (en) * | 2013-12-27 | 2015-07-01 | 比亚迪股份有限公司 | Lane deviation detection method and device |
| CN105160309A (en) * | 2015-08-24 | 2015-12-16 | 北京工业大学 | Three-lane detection method based on image morphological segmentation and region growing |
| CN105893949A (en) * | 2016-03-29 | 2016-08-24 | 西南交通大学 | Lane line detection method under complex road condition scene |
| CN106326822A (en) * | 2015-07-07 | 2017-01-11 | 北京易车互联信息技术有限公司 | Method and device for detecting lane line |
| CN106529493A (en) * | 2016-11-22 | 2017-03-22 | 北京联合大学 | Robust multi-lane line detection method based on perspective drawing |
| CN107025432A (en) * | 2017-02-28 | 2017-08-08 | 合肥工业大学 | A kind of efficient lane detection tracking and system |
| CN107066952A (en) * | 2017-03-15 | 2017-08-18 | 中山大学 | A kind of method for detecting lane lines |
-
2017
- 2017-11-03 CN CN201711070348.8A patent/CN107918763A/en active Pending
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004246641A (en) * | 2003-02-14 | 2004-09-02 | Nissan Motor Co Ltd | Traffic white lane line recognizing device |
| CN101608924A (en) * | 2009-05-20 | 2009-12-23 | 电子科技大学 | A Lane Line Detection Method Based on Gray Level Estimation and Cascaded Hough Transform |
| CN102324017A (en) * | 2011-06-09 | 2012-01-18 | 中国人民解放军国防科学技术大学 | A Lane Line Detection Method Based on FPGA |
| CN102303609A (en) * | 2011-06-16 | 2012-01-04 | 广东铁将军防盗设备有限公司 | Lane departure warning system and method |
| CN102360499A (en) * | 2011-06-30 | 2012-02-22 | 电子科技大学 | Multi-lane line tracking method based on Kalman filter bank |
| US20130202155A1 (en) * | 2012-02-03 | 2013-08-08 | Gopal Gudhur Karanam | Low-cost lane marker detection |
| CN103295420A (en) * | 2013-01-30 | 2013-09-11 | 吉林大学 | Method for recognizing lane line |
| CN104376297A (en) * | 2013-08-12 | 2015-02-25 | 株式会社理光 | Detection method and device for linear indication signs on road |
| CN104742912A (en) * | 2013-12-27 | 2015-07-01 | 比亚迪股份有限公司 | Lane deviation detection method and device |
| CN103971521A (en) * | 2014-05-19 | 2014-08-06 | 清华大学 | Method and device for detecting road traffic abnormal events in real time |
| CN106326822A (en) * | 2015-07-07 | 2017-01-11 | 北京易车互联信息技术有限公司 | Method and device for detecting lane line |
| CN105160309A (en) * | 2015-08-24 | 2015-12-16 | 北京工业大学 | Three-lane detection method based on image morphological segmentation and region growing |
| CN105893949A (en) * | 2016-03-29 | 2016-08-24 | 西南交通大学 | Lane line detection method under complex road condition scene |
| CN106529493A (en) * | 2016-11-22 | 2017-03-22 | 北京联合大学 | Robust multi-lane line detection method based on perspective drawing |
| CN107025432A (en) * | 2017-02-28 | 2017-08-08 | 合肥工业大学 | A kind of efficient lane detection tracking and system |
| CN107066952A (en) * | 2017-03-15 | 2017-08-18 | 中山大学 | A kind of method for detecting lane lines |
Non-Patent Citations (4)
| Title |
|---|
| ERKE SHANG ET AL.: "Lane Detection using Steerable Filters and FPGA-based Implementation", 《2011 SIXTH INTERNATIONAL CONFERENCE ON IMAGE AND GRAPHICS》 * |
| 刘亚群: "基于视觉的车道级定位算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 王家恩 等: "车辆辅助驾驶系统中的三车道检测算法", 《汽车工程》 * |
| 罗四维 等: "《计算机视觉检测逆问题导论》", 31 March 2017 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110838233A (en) * | 2019-10-12 | 2020-02-25 | 中国平安人寿保险股份有限公司 | Vehicle behavior analysis method and device and computer readable storage medium |
| CN110838233B (en) * | 2019-10-12 | 2022-09-27 | 中国平安人寿保险股份有限公司 | Vehicle behavior analysis method and device and computer readable storage medium |
| CN113822218A (en) * | 2021-09-30 | 2021-12-21 | 厦门汇利伟业科技有限公司 | Lane line detection method and computer-readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10452931B2 (en) | Processing method for distinguishing a three dimensional object from a two dimensional object using a vehicular system | |
| US10943131B2 (en) | Image based lane marking classification | |
| CN104573646B (en) | Chinese herbaceous peony pedestrian detection method and system based on laser radar and binocular camera | |
| EP3007099B1 (en) | Image recognition system for a vehicle and corresponding method | |
| US10891738B2 (en) | Boundary line recognition apparatus and branch road determination apparatus | |
| EP2958054B1 (en) | Hazard detection in a scene with moving shadows | |
| WO2012172713A1 (en) | Device for determining road profile, onboard image-recognition device, device for adjusting image-capturing axis, and lane-recognition method. | |
| US10457283B2 (en) | Vehicle driving assist apparatus | |
| CN112069944A (en) | Road congestion level determination method | |
| Kim | Realtime lane tracking of curved local road | |
| CN111149131B (en) | Dividing line recognition device | |
| Zhang et al. | Robust inverse perspective mapping based on vanishing point | |
| US10108866B2 (en) | Method and system for robust curb and bump detection from front or rear monocular cameras | |
| EP3046049A1 (en) | Complex marking determining device and complex marking determining method | |
| US20200125865A1 (en) | Roadside Object Detection Device, Roadside Object Detection Method, and Roadside Object Detection System | |
| Chen et al. | A novel lane departure warning system for improving road safety | |
| CN107918763A (en) | Method for detecting lane lines and system | |
| Takahashi et al. | A robust lane detection using real-time voting processor | |
| CN115880658A (en) | Automobile lane departure early warning method and system under night scene | |
| Lee et al. | A cumulative distribution function of edge direction for road-lane detection | |
| CN111881878B (en) | Lane line identification method for look-around multiplexing | |
| JP3319383B2 (en) | Roadway recognition device | |
| CN113345035A (en) | Binocular camera-based gradient real-time prediction method and system and computer-readable storage medium | |
| CN118545081A (en) | Lane departure warning method and system | |
| KR102418344B1 (en) | Traffic information analysis apparatus and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| AD01 | Patent right deemed abandoned | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20220913 |