A nighttime highway lane line detection method based on reflective road studs

Abstract

The present invention discloses a nighttime highway lane line detection method based on reflective road studs. Based on a high-quality nighttime front image acquisition system, setting a region of interest for lane line detection, performing image segmentation with an adaptive threshold, and extracting target pixel points of the reflective road studs, the reflective road stud pixel points in the region of interest are expanded using morphological operations. On this basis, left and right lane line fitting regions are divided according to the positions of bright blocks of the reflective road studs. Hough transform is used to perform straight line fitting in each of the left and right lane line fitting regions to extract the left and right lane lines. Safety thresholds for the slopes of the left and right lane lines are set. If the lane line slopes are not within the safety threshold range, a lane line deviation alarm is issued. This effectively improves the effectiveness of lane line detection in nighttime environments and enhances driving safety.

Description

Technical Field

[0001] The present invention relates to highway lane line detection in a nighttime environment, and in particular to a method for detecting highway lane lines at nighttime based on reflective road studs. Background Art

[0002] Lane detection is a key component of intelligent road transport safety systems and a key enabler for the safe and reliable operation of autonomous driving technology. Nighttime highway lane detection and departure warning are core technologies for driver assistance. Given the abundant daylight and the high contrast between lane lines and the surrounding environment, image processing techniques can be effectively used to extract clear lane lines. However, in low-light or complex nighttime lighting conditions, the image contrast between lane lines and the surrounding environment is insufficient, making effective lane detection difficult.

[0003] Currently, researchers at home and abroad are focusing on developing various illumination compensation methods to enhance the contrast of nighttime images for lane detection and recognition in nighttime environments. These methods, however, fail to address the issues of low image brightness and unclear lane features. Furthermore, they complicate lane detection and recognition, reducing the real-time performance of lane departure warnings. Therefore, a more accurate and efficient lane departure warning method for nighttime lane detection, applicable to various complex environments, is needed to assist drivers of all types in their travels, reduce the probability of accidents, and create a safer public transportation environment.

[0004] Given the current state of research in this field, in-depth analysis and consideration are needed for lane marking detection on highways at night. Considering the recent trend of installing reflective road studs on highway lanes to improve driver safety, reflective road studs reflect high brightness from vehicle headlights and can be utilized as a distinctive feature of lane markings. Summary of the Invention

[0005] The present invention aims to overcome the shortcomings of the prior art and provide a nighttime highway lane marking detection method based on reflective road studs. Based on a high-quality nighttime forward image acquisition system, setting a region of interest for lane marking detection, performing image segmentation using an adaptive threshold, and extracting target pixels of the reflective road studs, the reflective road stud pixels in the region of interest are then dilated using morphological operations. Furthermore, left and right lane marking fitting regions are divided according to the positions of bright blocks of the reflective road studs. Hough transforms are then used to perform straight line fitting within each of the left and right lane marking fitting regions to extract the left and right lane markings. Safety thresholds for the slopes of the left and right lane markings are set, and a lane deviation alarm is issued if the lane marking slopes are outside the safety threshold range. This effectively improves the effectiveness of lane marking detection in nighttime environments and enhances driving safety.

[0006] The object of the present invention is achieved through the following technical solutions:

[0007] A method for detecting lane markings on a highway at night based on reflective road studs comprises the following steps:

[0008] S1. Build a high-quality nighttime forward image acquisition system to obtain the image to be detected;

[0009] S2. Set the image region of interest to be detected, including the left and right lane marking areas;

[0010] S3. Segment the image to be detected using an adaptive threshold method to extract target pixels of the reflective road studs;

[0011] S4. dilate the reflective road stud pixels in the region of interest using a morphological operation;

[0012] S5. Divide the left and right lane line fitting areas based on the positions of the reflective road stud bright blocks;

[0013] S6. Perform straight line fitting using Hough transform in the left and right lane marking regions, thereby obtaining the left and right lane markings;

[0014] S7. Extract the slopes of the left and right lane lines respectively, set safety thresholds for the slopes of the left and right lane lines, and issue a lane departure alarm if the lane line slopes are not within the safety threshold range.

[0015] A further improvement of the present invention is that the specific implementation method of step S1 is as follows:

[0016] Taking into account the image quality and stability of the detection system, an industrial camera is selected as the image acquisition module of the system; it is fixed to the front windshield with a suction cup, and the camera housing is set to be adjustable in up and down pitch angles and left and right yaw angles; the camera lens uses a variable focus lens that can adjust the brightness and color saturation of the captured image.

[0017] A further improvement of the present invention is that the specific implementation method of step S2 is as follows:

[0018] Select any image containing left and right lane lines as a sample image, and select the area of ​​the image containing the left and right lane lines as a template; mark the position coordinates of the four vertices of the rectangular box where the lane line area is located, and scale it according to the calibration position to obtain the detection area of ​​interest for the lane line.

[0019] A further improvement of the present invention is that the specific implementation method of step S3 is as follows:

[0020] S301. Acquire several images containing reflective road studs and lane markings. The original images are RGB images composed of the three primary colors red (R), green (G), and blue (B). Convert the original RGB images to the HSV color model. The conversion process is as follows:

[0021] R′=R / 255 (1)

[0022] G′=G / 255 (2)

[0023] B′=B / 255 (3)

[0024] C max =max(R′,G′,B′) (4)

[0025] C min =min(R′,G′,B′) (5)

[0026] Δ=C max -C min (6)

[0027]

[0028]

[0029] V=C max (9)

[0030] S302. The HSV color image obtained in step S301 is subjected to three-channel separation processing to obtain grayscale images of the three components H, S, and V. The histograms of the grayscale images of the three components are synthesized to obtain a total histogram; the total histogram obtained above is equalized using a limited contrast histogram equalization algorithm;

[0031] S303 uses the median filter method to denoise, separates the image in the HSV color mode into three channels, and performs a median filter on all pixels of each channel image in a rectangular full search mode, that is, the value of each pixel is calculated by accumulating the median value of its neighboring pixels;

[0032] S304. Based on S301 to S303, extract the brightness component V of HSV and segment V;

[0033] S305. In the V component, a large number of reflective road stud bright block regions are captured, and the V component histograms of these regions are calculated. The segmentation threshold range for the reflective road stud bright blocks is determined to be [m1, m2], where m1 is the minimum threshold and m2 is the maximum threshold. The V component images are segmented using an improved maximum inter-class variance method, where the grayscale search range for the maximum inter-class variance is reduced from [0, 255] to [m1, m2]. This segmentation yields a binary image with the reflective road stud bright blocks as the target.

[0034] A further improvement of the present invention is that the specific implementation method of step S4 is as follows:

[0035] S401. In the target binary image obtained from S301 to S305, the bright block area of ​​the reflective road stud often has a small hole at the center. Morphological processing is performed on the binary image to fill the hole. Specifically, a point in the hole is first extracted, then continuously expanded using a structuring element, and then the intersection is calculated with the complement of the original image. This process is repeated until the image no longer changes, and finally the intersection is calculated with the original image.

[0036] S402. Obtain the target binary map from S301 to S305, connect target pixels and annotate them, and extract attributes such as the area, perimeter, centroid position, and minimum bounding rectangle coordinates of each connected region. These connected regions are designated as candidate target regions for reflective road studs.

[0037] A further improvement of the present invention is that the specific implementation method of step S5 is as follows:

[0038] S501. The position of the circumscribed rectangular frame of the reflective road stud bright block obtained from S402 above is extracted, and the X-axis coordinate of the upper left corner is sorted from small to large to mark the positional relationship of the reflective road stud bright block;

[0039] S502. To prevent interference from reflective stickers on highway guardrails, streetlights, and billboards on the final straight-line fitting results for the reflective road studs, strict left and right lane fitting regions are defined along the straight-line fitting line. Dataset analysis and extensive experiments have shown that the vast majority of reflective road studs fall within a triangular region. Connecting the vertex and the midpoint of the base, the large triangle is then divided into two smaller left and right triangles. These two smaller triangles constitute the left and right lane detection regions.

[0040] A further improvement of the present invention is that the specific implementation method of step S6 is as follows:

[0041] S601. First, a random edge point is extracted from the image, the edge point is marked as a point on the line, and the next edge point is randomly extracted until all edge points are extracted;

[0042] S602. After performing Hough transform on the randomly extracted edge points, the cumulative sum is calculated; the maximum point of the accumulator is selected, and if it is higher than the set threshold, the process proceeds to S603, otherwise it returns to S601;

[0043] S603. Starting from the selected point, find the two endpoints of the line along the direction of linear displacement; calculate the length of the line segment. If it is greater than the set threshold, the line segment is determined to be a lane line. The loop starts by calculating the cumulative sum of the Hough transform of the edge points to obtain the left and right lane lines.

[0044] A further improvement of the present invention is that the specific implementation method of step S7 is as follows:

[0045] S701. Set the slope safety threshold of the left lane line to -0.43. When the slope of the left lane line detected in real time is greater than the threshold, a left vehicle wire deviation alarm is issued;

[0046] S702. Set the right lane line slope safety threshold to 0.8, and issue a right lane line deviation alarm when the real-time detection of the right lane line slope is greater than the threshold;

[0047] This enables the nighttime highway lane line detection function based on reflective road studs.

[0048] The present invention has at least the following beneficial technical effects:

[0049] This paper proposes a method for detecting lane markings in nighttime environments using reflective road studs as lane marking features. As target features for detecting lane markings on highways at night, reflective road studs offer high brightness, distinctness from the background, regular shapes, and uniform distribution, meeting the requirements for lane marking detection based on reflective road studs in nighttime environments.

[0050] Furthermore, a high-quality nighttime forward image acquisition system was built. Image enhancement preprocessing was performed on the images to be identified to improve image quality. A region of interest for lane line detection was set to effectively reduce interfering bright blocks, thereby improving the accuracy and speed of the detection method. Adaptive thresholding was used for image segmentation and morphological processing was used to extract reflective road stud bright blocks. On this basis, the distribution positions of reflective road stud bright blocks were statistically analyzed based on a large number of experimental results, and left and right lane line fitting regions were divided. Hough transform was used to fit straight lines to the left and right lane line fitting regions, respectively. Finally, safety thresholds for the slopes of the left and right lane lines were set. If the lane line slopes were not within the safety threshold range, a lane line deviation alarm was issued. This effectively solved the problems of missed and false detections of lane lines due to insufficient lighting in nighttime environments, improved the accuracy of lane line detection and deviation alarms at night, and had great engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a schematic diagram of an algorithm flow chart of an embodiment of the present invention;

[0052] Figure 2 is a schematic diagram of setting a region of interest according to an embodiment of the present invention;

[0053] Figure 3 is a schematic diagram of a brightness component V according to an embodiment of the present invention;

[0054] Figure 4 is a binary image after segmentation according to an embodiment of the present invention;

[0055] Figure 5 is a morphological processing effect diagram of an embodiment of the present invention; wherein Figure 5 (a) is the edge extraction result diagram, Figure 5 (b) is the morphological processing effect diagram;

[0056] Figure 6 This is a schematic diagram of the left and right lane detection areas according to an embodiment of the present invention;

[0057] Figure 7 This is a rendering of the left lane line extraction effect according to an embodiment of the present invention;

[0058] Figure 8 This is a rendering of the right lane line extraction effect according to an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0060] The present invention provides a method for detecting lane lines on highways at night based on reflective road studs. The algorithm flow chart is as follows: Figure 1 shown.

[0061] The following steps are involved:

[0062] S1. Build a high-quality nighttime forward image acquisition system;

[0063] S2. Set the image region of interest to be detected, including the left and right lane marking areas;

[0064] S3. Segment the image to be detected using an adaptive threshold method to extract target pixels of the reflective road studs;

[0065] S4. dilate the reflective road stud pixels in the region of interest using a morphological operation;

[0066] S5. Divide the left and right lane line fitting areas based on the positions of the reflective road stud bright blocks;

[0067] S6. Perform straight line fitting using Hough transform in the left and right lane marking regions, thereby obtaining the left and right lane markings;

[0068] S7. Extract the slopes of the left and right lane lines respectively, set safety thresholds for the slopes of the left and right lane lines, and issue a lane departure alarm if the lane line slopes are not within the safety threshold range.

[0069] In an embodiment of the present application, step S1 includes:

[0070] Considering image quality and system stability, an industrial camera was selected as the system's image acquisition module. It was fixed to the front windshield using a suction cup, and the camera housing was designed to be adjustable for both vertical and horizontal tilt. A variable-focus lens was used to flexibly adjust the brightness and color saturation of the captured image. This mounting method offers the advantage of adjustable camera position, making image quality adjustments flexible and simple.

[0071] In an embodiment of the present application, step S2 includes:

[0072] Select any image containing left and right lane lines as a sample image, and select the area containing left and right lane lines as a template. Mark the position coordinates of the four vertices of the rectangular box where the lane line area is located, and scale it according to the calibration position to obtain the detection area of ​​interest for the lane line. The diagram of the area of ​​interest setting is as follows: Figure 2 shown.

[0073] In an embodiment of the present application, step S3 includes:

[0074] S301. Acquire several images containing reflective road studs and lane markings. The original images are RGB images, composed of the three primary colors red (R), green (G), and blue (B). Extensive experiments have shown that converting the original images to the HSV color model produces greater contrast in the reflective road studs, facilitating subsequent extraction of the reflective road stud region of interest. Therefore, the original RGB images are converted to the HSV color model. The conversion process is shown below:

[0075] R′=R / 255 (1)

[0076] G′=G / 255 (2)

[0077] B′=B / 255 (3)

[0078] C max =max(R′,G′,B′) (4)

[0079] Cmin =min(R′,G′,B′) (5)

[0080] Δ=C max -C min (6)

[0081]

[0082]

[0083] V=C max (9)

[0084] S302. The HSV color image obtained in step S301 is subjected to three-channel separation, resulting in grayscale images of the H, S, and V components. The histograms of these three components are then combined to form a total histogram. The total histogram obtained is then equalized using the Contrast Constrained Histogram Equalization (CLAHE) algorithm. This method effectively improves local image contrast and captures more image detail while suppressing noise enhancement.

[0085] S303. During image acquisition and transmission, the captured images are often subject to high levels of noise due to the surrounding environment. This can severely result in the loss or alteration of certain image features, hindering the implementation of the reflective stud and lane marking recognition method. To address this issue, a median filter is used for denoising. The HSV color image is separated into three channels. A median filter is then applied to all pixels in each channel using a full rectangular search. This means that the value of each pixel is calculated by accumulating the median value of its neighboring pixels.

[0086] S304. Based on S301 to S303, extract the brightness component V of HSV, as shown in the schematic diagram Figure 3 As shown, V is segmented;

[0087] S305. In the V component, a large number of reflective road stud bright block areas are intercepted from the image, and the V component histogram of such areas is calculated. The segmentation threshold range for the reflective road stud bright block is obtained as [m1, m2], where m1 is the minimum threshold and m2 is the maximum threshold. The V component image is segmented using the improved maximum inter-class variance method, that is, the grayscale search range of the maximum inter-class variance is reduced from [0, 255] to [m1, m2]. After segmentation, a binary image with the reflective road stud bright block as the target is obtained. The binary image is as follows: Figure 4 As shown;

[0088] In an embodiment of the present application, step S4 includes:

[0089] S401. In the target binary image obtained from S301 to S305, the bright block area of ​​the reflective road stud often has a small hole in the center. Morphological processing is performed on the binary image to fill the hole. The morphological processing effect is shown in the figure below. Figure 5 As shown in the figure, the method is: first extract a point in the hole, then continuously dilate it with the structure element, then intersect it with the complement of the original image, repeat this process until the image no longer changes, and finally intersect it with the original image;

[0090] S402. Obtain the target binary map from S301 to S305, connect target pixels and annotate them, and extract attributes such as the area, perimeter, centroid position, and minimum bounding rectangle coordinates of each connected region. These connected regions are designated as candidate target regions for reflective road studs.

[0091] In an embodiment of the present application, step S5 includes:

[0092] S501. The position of the circumscribed rectangular frame of the reflective road stud bright block obtained from S402 above is extracted, and the X-axis coordinate of the upper left corner is sorted from small to large to mark the positional relationship of the reflective road stud bright block;

[0093] S502. To prevent interference from reflective stickers on highway guardrails, streetlights, billboards, and other interfering bright blocks from affecting the final straight-line fitting results for reflective road studs, a strict left lane line fitting area and a right lane line fitting area are divided along the straight-line fitting line. Through data set analysis and extensive experiments, it was found that the vast majority of reflective road studs fall within a triangular area. Connecting the vertex of the triangle to the midpoint of the base, the large triangle is then divided into two smaller triangles on the left and right. These two smaller triangles represent the left and right lane line detection areas, as shown in the diagram below. Figure 6 shown.

[0094] In an embodiment of the present application, step S6 includes:

[0095] S601. First, a random edge point is extracted from the image, the edge point is marked as a point on the line, and the next edge point is randomly extracted until all edge points are extracted;

[0096] S602. Perform Hough transform on the randomly extracted edge points and then perform cumulative sum calculation. Select the maximum point of the accumulator and proceed to S603 if it is higher than the set threshold, otherwise return to S601;

[0097] S603. Start from the selected point and find the two endpoints of the straight line along the direction of the straight line displacement. Calculate the length of the line segment. If it is greater than the set threshold, the line segment is determined to be a lane line. The cycle starts from the accumulation and calculation of the edge point after Hough transformation, thereby obtaining the left and right lane lines. The left and right lane line extraction effect is shown in the figure below. Figure 7 and Figure 8 shown.

[0098] In an embodiment of the present application, step S7 includes:

[0099] S701. Set the slope safety threshold of the left lane line to -0.43. When the slope of the left lane line detected in real time is greater than the threshold, a left vehicle wire deviation alarm is issued;

[0100] S702. Set the slope safety threshold of the right lane line to 0.8. When the slope of the right lane line detected in real time is greater than the threshold, a right lane line deviation alarm is issued.

[0101] This enables the nighttime highway lane line detection function based on reflective road studs.

[0102] The foregoing description shows and describes a preferred embodiment of the present invention. However, as previously stated, it should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Instead, the present invention is applicable to various other combinations, modifications, and environments and is capable of modification within the scope of the inventive concept described herein, through the teachings above, or through techniques or knowledge in the relevant art. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be within the scope of the appended claims.

Claims

1. A method for detecting lane lines on highways at night based on reflective road studs, characterized in that: The following steps are involved: S1. Build a high-quality nighttime forward image acquisition system to obtain the image to be detected; S2. Set the image region of interest to be detected, including the left and right lane marking areas; S3. Segment the image to be detected using an adaptive threshold method to extract the target pixels of the reflective road studs. The specific implementation method is as follows: S301. Acquire several images containing reflective road studs and lane markings. The original images are RGB images composed of the three primary colors red (R), green (G), and blue (B). Convert the original RGB images to the HSV color model. The conversion process is as follows: R′=R / 255 (1) G′=G / 255 (2) B′=B / 255 (3) C max =max(R′,G′,B′) (4) C min =min(R′,G′,B′) (5) Δ=C max -C min (6) V=C max (9) S302. The HSV color image obtained in step S301 is subjected to three-channel separation processing to obtain grayscale images of the three components H, S, and V. The histograms of the grayscale images of the three components are synthesized to obtain a total histogram; The total histogram obtained above is equalized using the limited contrast histogram equalization algorithm; S303 uses the median filter method to denoise, separates the image in the HSV color mode into three channels, and performs a median filter on all pixels of each channel image in a rectangular full search mode, that is, the value of each pixel is calculated by accumulating the median value of its neighboring pixels; S304. Based on S301 to S303, extract the brightness component V of HSV and segment V; S305. In the V component, a large number of reflective road stud bright block regions are captured, and the V component histograms of these regions are calculated. The segmentation threshold range for the reflective road stud bright blocks is determined to be [m1, m2], where m1 is the minimum threshold and m2 is the maximum threshold. The V component image is segmented using an improved maximum inter-class variance method, where the grayscale search range for the maximum inter-class variance is reduced from [0, 255] to [m1, m2]. This segmentation yields a binary image with the reflective road stud bright blocks as the target. S4. dilate the reflective road stud pixels in the region of interest using a morphological operation; S5. Divide the left and right lane line fitting areas based on the positions of the reflective road stud bright blocks; S6. Perform straight line fitting using Hough transform in the left and right lane marking regions, thereby obtaining the left and right lane markings; S7. Extract the slopes of the left and right lane lines respectively, set safety thresholds for the slopes of the left and right lane lines, and issue a lane departure alarm if the lane line slopes are not within the safety threshold range.

2. The method for detecting lane lines on highways at night based on reflective road studs according to claim 1, characterized in that: The specific implementation method of step S1 is as follows: Taking into account the image quality and stability of the detection system, an industrial camera is selected as the image acquisition module of the system; it is fixed to the front windshield with a suction cup, and the camera housing is set to be adjustable in up and down pitch angles and left and right yaw angles; the camera lens uses a variable focus lens that can adjust the brightness and color saturation of the captured image.

3. The method for detecting lane lines on highways at night based on reflective road studs according to claim 1, characterized in that: The specific implementation method of step S2 is as follows: Select any image containing left and right lane lines as a sample image, and select the area of ​​the image containing the left and right lane lines as a template; mark the position coordinates of the four vertices of the rectangular box where the lane line area is located, and scale it according to the calibration position to obtain the detection area of ​​interest for the lane line.

4. The method for detecting lane lines on highways at night based on reflective road studs according to claim 1, characterized in that: The specific implementation method of step S4 is as follows: S401. In the target binary image obtained from S301 to S305, the bright block area of ​​the reflective road stud often has a small hole at the center. Morphological processing is performed on the binary image to fill the hole. Specifically, a point in the hole is first extracted, then continuously expanded using a structuring element, and then the intersection is calculated with the complement of the original image. This process is repeated until the image no longer changes, and finally the intersection is calculated with the original image. S402. Obtain the target binary map from S301 to S305, connect target pixels and annotate them, and extract attributes such as the area, perimeter, centroid position, and minimum bounding rectangle coordinates of each connected region. These connected regions are designated as candidate target regions for reflective road studs.

5. The method for detecting lane lines on highways at night based on reflective road studs according to claim 4, characterized in that: The specific implementation method of step S5 is as follows: S501. The position of the circumscribed rectangular frame of the reflective road stud bright block obtained from S402 above is extracted, and the X-axis coordinate of the upper left corner is sorted from small to large to mark the positional relationship of the reflective road stud bright block; S502. To prevent interference from reflective stickers on highway guardrails, streetlights, and billboards on the final straight-line fitting results for the reflective road studs, strict left and right lane fitting regions are defined along the straight-line fitting line. Dataset analysis and extensive experiments have shown that the vast majority of reflective road studs fall within a triangular region. Connecting the vertex and the midpoint of the base, the large triangle is then divided into two smaller left and right triangles. These two smaller triangles constitute the left and right lane detection regions.

6. The method for detecting lane lines on highways at night based on reflective road studs according to claim 1, characterized in that: The specific implementation method of step S6 is as follows: S601. First, a random edge point is extracted from the image, the edge point is marked as a point on the line, and the next edge point is randomly extracted until all edge points are extracted; S602. After performing Hough transform on the randomly extracted edge points, the cumulative sum is calculated; the maximum point of the accumulator is selected, and if it is higher than the set threshold, the process proceeds to S603, otherwise it returns to S601; S603. Starting from the selected point, find the two endpoints of the line along the direction of linear displacement; calculate the length of the line segment. If it is greater than the set threshold, the line segment is determined to be a lane line. The loop starts by calculating the cumulative sum of the Hough transform of the edge points to obtain the left and right lane lines.

7. The method for detecting lane lines on highways at night based on reflective road studs according to claim 1, characterized in that: The specific implementation method of step S7 is as follows: S701. Set the slope safety threshold of the left lane line to -0.

43. When the slope of the left lane line detected in real time is greater than the threshold, a left vehicle wire deviation alarm is issued; S702. Set the right lane line slope safety threshold to 0.8, and issue a right lane line deviation alarm when the real-time detection of the right lane line slope is greater than the threshold; This enables the nighttime highway lane line detection function based on reflective road studs.

Images