CN108052904B - Method and device for acquiring lane line - Google Patents

Abstract

The present invention provides a method and device for acquiring lane lines. The method includes: performing inverse perspective transformation on a region of interest map on a road to be detected to obtain a transformed region of interest map; and determining the transformation by using a YC _b C _r color space model. The color probability of each pixel belonging to the lane line in the area of interest map after the The gray probability belonging to the lane line is used to obtain the gray probability map of the lane line; the clustering algorithm is used to segment the gray probability model map of the lane line to obtain a binary segmentation result map; the binary segmentation result map is processed and obtained Lane lines in the road to be detected. The method and device for acquiring lane lines provided by the present invention improve the detection accuracy of lane lines.

Description

Method and device for obtaining lane lines

technical field

The invention relates to the technical field of intelligent transportation, and in particular, to a method and device for acquiring lane lines.

Background technique

Automotive assisted driving systems are often used to avoid driver errors and fatigue driving. When the vehicle deviates from the correct lane or is about to collide, it reminds the driver or automatically controls the vehicle to a safe state, thereby improving driver safety. In the vehicle assisted driving system, the accuracy and robustness of lane line detection are very important.

In order to improve the detection accuracy of the lane line, in the prior art, the method based on the color feature of the lane line and the method based on the gray value feature of the edge of the lane line are usually used for detection. However, for the method based on the color feature of the lane line, its It is required that the lane lines must have sharp contrasting colors, and the existing lane lines do not necessarily have sharp contrasts between the colors, so the detection accuracy of the lane lines using this method is not high. In addition, for the method based on the feature of the gray value of the edge of the lane line, if the similarity between the features of the gray value of the edge of the lane line is high on the road with many safety marks, the accuracy of the detection of the lane line by this method is not good. high.

Therefore, by using the existing lane line detection method, the accuracy of lane line detection is not high.

SUMMARY OF THE INVENTION

The present invention provides a method and device for acquiring lane lines, so as to improve the detection accuracy of lane lines.

An embodiment of the present invention provides a method for acquiring lane lines, including:

Perform inverse perspective transformation on the ROI map in the road to be detected to obtain the transformed ROI map;

Using the YC _b C _r color space model to determine the color probability that each pixel in the transformed region of interest map belongs to the lane line;

and normalize the color probability that each pixel belongs to the lane line to obtain the grayscale probability that each pixel belongs to the lane line;

Obtain the lane line gray level probability map according to the gray level probability that each pixel belongs to the lane line;

A clustering algorithm is used to perform regional segmentation processing on the lane line grayscale probability model graph to obtain a binary segmentation result graph;

The binary segmentation result graph is processed, and lane lines in the road to be detected are acquired.

In an embodiment of the present invention, the use of the YC _b C _r color space model to determine the color probability of each pixel belonging to the lane line in the transformed region of interest map includes:

确定所述变换后的感兴趣区域图中每一个像素属于车道线的颜色概率；according to

Determine the color probability that each pixel in the transformed region of interest map belongs to the lane line;

为先验性概率，#C_i为颜色C_i类中的样本数，K表示第K种颜色。Among them, P(C _i |x) represents the probability that the pixel belongs to the lane line, N-1 represents that there are N-1 types of colors in the lane line, and C _i is the color of the ith type. The color of each pixel can be expressed as: x=( Y, C _b , C _r ), and x=(Y, C _b , C _r ), Y is the luminance component, C _b is the blue chrominance component, C _r is the red chrominance component, and P(x) is the pixel x The color probability of , P(x|C _i ) is the probability probability,

is the prior probability, #C _i is the number of samples in the color C _i class, and K represents the Kth color.

In an embodiment of the present invention, the normalization process is performed on the color probability that each pixel belongs to the lane line to obtain the grayscale probability that each pixel belongs to the lane line, including:

对所述每一个像素属于车道线的颜色概率进行归一化处理，得到所述每一个像素属于车道线的灰度概率；according to

Normalize the color probability that each pixel belongs to the lane line to obtain the grayscale probability that each pixel belongs to the lane line;

表示像素属于车道线的灰度概率。in,

Represents the grayscale probability that a pixel belongs to a lane line.

In an embodiment of the present invention, the processing of the binary segmentation result map and the acquisition of lane lines in the road to be detected include:

Morphological processing is performed on the binary segmentation result map to obtain a processed binary segmentation result map;

The processed binary segmentation result graph is processed, and lane lines in the road to be detected are acquired.

In an embodiment of the present invention, the processing of the processed binary segmentation result graph, and the acquisition of lane lines in the road to be detected, includes:

Adopt Sobel Sobel algorithm to detect the edge curve in the binary segmentation result figure after described processing;

Perform centerline operation processing on the edge curve to obtain the processed center curve;

The center curve is segmented by using Hall Hough transform, and a plurality of straight line segments obtained after segmental processing are fitted to obtain lane lines in the road to be detected.

In an embodiment of the present invention, the Sobel algorithm is used to detect the edge curve in the processed binary segmentation result graph, including:

Adopt the described Sobel algorithm to obtain the horizontal value and the vertical value of each pixel in the processed binary segmentation result graph;

Determine the gradient value and direction of each pixel according to the horizontal value and the vertical value of each pixel;

When the gradient value and direction of each pixel are greater than a preset threshold, and the gradient value is within a certain preset range, determine that each pixel is an edge point of the processed binary segmentation result map;

The edge curve in the processed binary segmentation result graph is detected according to each edge point.

In an embodiment of the present invention, after acquiring the lane lines in the road to be detected, the method includes:

The lane lines in the road to be detected are displayed on the map.

An embodiment of the present invention also provides a device for acquiring lane lines, including:

The transformation unit is used to perform inverse perspective transformation on the ROI map in the road to be detected to obtain the transformed ROI map;

A determination unit, used for using the YC _b C _r color space model to determine the color probability that each pixel in the transformed region of interest map belongs to the lane line;

a processing unit, configured to perform normalization processing on the color probability that each pixel belongs to the lane line, to obtain the grayscale probability that each pixel belongs to the lane line;

a generating unit, configured to obtain the grayscale probability map of the lane line according to the grayscale probability that each pixel belongs to the lane line;

a segmentation unit, used for using a clustering algorithm to perform regional segmentation processing on the lane line grayscale probability model graph to obtain a binary segmentation result graph;

an acquisition unit, configured to process the binary segmentation result map and acquire lane lines in the road to be detected.

N确定所述变换后的感兴趣区域图中每一个像素属于车道线的颜色概率；In an embodiment of the present invention, the determining unit is specifically configured to

N determines the color probability that each pixel in the transformed region of interest map belongs to the lane line;

为先验性概率，#C_i为颜色C_i类中的样本数，K表示第K种颜色。Among them, P(C _i |x) represents the probability that the pixel belongs to the lane line, N-1 represents that there are N-1 types of colors in the lane line, and C _i is the color of the ith type. The color of each pixel can be expressed as: x=( Y, C _b , C _r ), and x=(Y, C _b , C _r ), Y is the luminance component, C _b is the blue chrominance component, C _r is the red chrominance component, and P(x) is the pixel x The color probability of , P(xC _i ) is the probability probability,

is the prior probability, #C _i is the number of samples in the color C _i class, and K represents the Kth color.

对所述每一个像素属于车道线的颜色概率进行归一化处理，得到所述每一个像素属于车道线的灰度概率；In an embodiment of the present invention, the processing unit is specifically configured to

Normalize the color probability that each pixel belongs to the lane line to obtain the grayscale probability that each pixel belongs to the lane line;

表示像素属于车道线的灰度概率。in,

Represents the grayscale probability that a pixel belongs to a lane line.

In an embodiment of the present invention, the acquisition unit is specifically configured to perform morphological processing on the binary segmentation result map to obtain a processed binary segmentation result map; and perform morphological processing on the processed binary segmentation result. image processing, and obtain the lane lines in the road to be detected.

In an embodiment of the present invention, the obtaining unit is specifically configured to use the Sobel algorithm to detect the edge curve in the processed binary segmentation result graph; perform centerline operation processing on the edge curve, and obtain the processed The center curve is obtained; the center curve is segmented by using Hall Hough transform, and a plurality of straight line segments obtained after segmental processing are fitted to obtain the lane lines in the road to be detected.

In an embodiment of the present invention, the obtaining unit is specifically configured to use the Sobel algorithm to obtain the horizontal value and vertical value of each pixel in the processed binary segmentation result graph; The horizontal value and the vertical value determine the gradient value and direction of each pixel; when the gradient value and direction of each pixel are greater than a preset threshold, and the gradient value is within a certain preset range, determine that each pixel is The edge point of the processed binary segmentation result graph; the edge curve in the processed binary segmentation result graph is detected according to each edge point.

In an embodiment of the present invention, the device for acquiring lane lines further includes a display unit;

The display unit is used for displaying lane lines in the road to be detected on a map.

An embodiment of the present invention provides a method for acquiring lane lines. First, inverse perspective transformation is performed on the ROI map on the road to be detected to obtain the transformed ROI map; the YC _b C _r color space model is used to determine the transformed ROI map. The color probability of each pixel belonging to the lane line in the area of interest map; and the color probability of each pixel belonging to the lane line is normalized to obtain the grayscale probability of each pixel belonging to the lane line; according to the fact that each pixel belongs to the lane The gray probability of the line is used to obtain the gray probability map of the lane line; the clustering algorithm is used to segment the gray probability model map of the lane line to obtain a binary segmentation result map; Lane lines in the road, thereby improving the accuracy of lane line detection.

Description of drawings

FIG. 1 is a schematic diagram of a method for acquiring lane lines according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a camera position according to an embodiment of the present invention;

3 is a schematic diagram of the size of a captured frame image according to an embodiment of the present invention;

4 is a schematic diagram of a region of interest map before and after inverse perspective transformation of a region of interest map provided by an embodiment of the present invention;

5 is a schematic diagram of a binary segmentation result graph provided by an embodiment of the present invention;

6 is a schematic flowchart of processing a binary segmentation result graph according to an embodiment of the present invention;

7 is a schematic diagram of a binary segmentation result graph after morphological processing according to an embodiment of the present invention;

8 is a schematic diagram of a processed center curve according to an embodiment of the present invention;

9 is a schematic diagram of lane lines on both sides obtained after tracking according to an embodiment of the present invention;

10 is a schematic structural diagram of a device for obtaining lane lines according to an embodiment of the present invention;

FIG. 11 is a schematic structural diagram of another apparatus for obtaining lane lines according to an embodiment of the present invention.

Detailed ways

Exemplary embodiments will be described herein, examples of which are illustrated in the accompanying drawings. Where the following description refers to the drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the illustrative examples below are not intended to represent all implementations consistent with this disclosure.

The terms "first", "second", "third", "fourth", etc. (if present) in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to Describe a particular order or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the invention described herein can, for example, be practiced in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having" and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those expressly listed Rather, those steps or units may include other steps or units not expressly listed or inherent to these processes, methods, products or devices.

In the prior art, the detection method is usually based on the color feature of the lane line and the method based on the gray value feature of the edge of the lane line. However, for the method based on the color feature of the lane line, the lane line must have a sharp contrasting color. , and the colors of the existing lane lines do not necessarily have a sharp contrast, so the detection accuracy of the lane lines using this method is not high. In addition, for the method based on the feature of the gray value of the edge of the lane line, if the similarity between the features of the gray value of the edge of the lane line is high on the road with many safety marks, the accuracy of the detection of the lane line by this method is not good. high, resulting in low accuracy of lane line detection. In order to improve the accuracy of lane line detection, an embodiment of the present invention provides a method for acquiring lane lines. First, perform inverse perspective transformation on the ROI map on the road to be detected to obtain the transformed ROI map; use YC _b C The _r color space model determines the color probability of each pixel belonging to the lane line in the transformed region of interest map; and normalizes the color probability of each pixel belonging to the lane line to obtain the gray level of each pixel belonging to the lane line Probability; according to the gray probability that each pixel belongs to the lane line, the gray probability map of the lane line is obtained; the clustering algorithm is used to perform regional segmentation processing on the gray probability model map of the lane line, and the binary segmentation result map is obtained; Graphs are processed, and lane lines in the road to be detected are obtained, thereby improving the accuracy of lane line detection.

The technical solutions of the present invention will be described below with specific embodiments. The following specific embodiments may be combined with each other, and the same or similar concepts or processes will not be repeated in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a schematic diagram of a method for acquiring lane lines according to an embodiment of the present invention. The method for acquiring lane lines may be obtained by a device for acquiring lane lines, which may be set independently or integrated in other equipment. , please refer to Figure 1, the acquisition method of the lane line can include:

S101. Perform inverse perspective transformation on the ROI map on the road to be detected to obtain a transformed ROI map.

After the original image of the road to be detected is determined, the sky area in the region of interest in the original image can be removed to obtain the region of interest map; then inverse perspective transformation is performed on the region of interest, for example, when inverse perspective transformation is performed , the lane lines can be transformed from the camera view to a bird's-eye view by using the camera parameters, the lane lines after inverse perspective transformation are parallel and have the same width, and then the lanes are detected using filters or geometric constraints. The road map in the image coordinate system can be transformed into the world coordinate system using the inverse perspective transformation. 2 and FIG. 3, FIG. 2 is a schematic diagram of a camera position according to an embodiment of the present invention, and FIG. 3 is a schematic diagram of the size of a captured frame image according to an embodiment of the present invention. The coordinates of the region of interest after inverse perspective transformation are x, y, z, γ and θ are the yaw and pitch angles of the camera, and the camera's viewing angle range is 2α. The position of the camera relative to the ground plane is d, h, l, the coordinates in the initial frame image are u, v, and the size of the initial frame image is R _x , R _y .

Among them, the model of inverse perspective transformation can be:

z=0

After inverse perspective transformation is performed on the region of interest map, an inverse perspective transformation region of interest map can be generated. Please refer to FIG. 4 . FIG. 4 is a region of interest map provided by an embodiment of the present invention before inverse perspective transformation is performed. and a schematic representation of the transformed region of interest map.

S102 , using the YC _b C _r color space model to determine the color probability that each pixel in the transformed region of interest map belongs to the lane line.

After obtaining the transformed region of interest map through S101, the YC _b C _r color space model can be used to determine the color probability of each pixel belonging to the lane line in the inversely transformed region of interest map. The YC _b C _r space has better performance than the RGB space, and can reduce the storage capacity of the color image collected by the camera based on the characteristics of human vision. Therefore, the area of interest map after inverse perspective transformation can be converted into RGB color The space is converted to the YC _b C _r space, where Y is the luminance component, C _b is the blue chrominance component, and C _r is the red chrominance component, so that the transformed region of interest is determined by the YC _b C _r color space model in the map The color probability that each pixel belongs to the lane line.

RGB颜色空间转换为YC_bC_r空间。Specifically, through

RGB color space is converted to YC _b C _r space.

Optionally, in this embodiment of the present invention, S102 adopts the YC _b C _r color space model to determine the color probability of each pixel belonging to the lane line in the transformed region of interest map, including:

确定变换后的感兴趣区域图中每一个像素属于车道线的颜色概率。according to

Determine the color probability of each pixel belonging to the lane line in the transformed region of interest map.

为可能性概率，

为先验性概率，其中#C_i为颜色C_i类中的样本数，K表示第K种颜色。Among them, P(C _i |x) represents the probability that the pixel belongs to the lane line, N-1 represents that there are N-1 types of colors in the lane line, and C _i is the color of the ith type. The color of each pixel can be expressed as: x=( Y, C _b , C _r ), and x=(Y, C _b , C _r ), Y is the luminance component, C _b is the blue chrominance component, C _r is the red chrominance component, and P(x) is the pixel x The color probability of ,

is the probability probability,

is the prior probability, where #C _i is the number of samples in the color C _i class, and K represents the Kth color.

S103 , normalize the color probability of each pixel belonging to the lane line, and obtain the grayscale probability that each pixel belongs to the lane line.

After determining the color probability of each pixel belonging to the lane line in the transformed region of interest map through S102, the color probability of each pixel belonging to the lane line can be normalized to obtain the gray level of each pixel belonging to the lane line probability. Optionally, S103 normalizes the color probability that each pixel belongs to the lane line, and obtains the grayscale probability that each pixel belongs to the lane line, including:

对每一个像素属于车道线的颜色概率进行归一化处理，得到每一个像素属于车道线的灰度概率；according to

Normalize the color probability that each pixel belongs to the lane line, and obtain the gray probability that each pixel belongs to the lane line;

表示像素属于车道线的灰度概率。in,

Represents the grayscale probability that a pixel belongs to a lane line.

S104, obtaining a lane line gray probability map according to the gray probability that each pixel belongs to the lane line.

After normalizing the color probability of each pixel belonging to the lane line through S103 to obtain the grayscale probability of each pixel belonging to the lane line, the grayscale of the lane line can be obtained according to the grayscale probability of each pixel belonging to the lane line Probability map, in this grayscale probability map of lane lines, the lane line area has a higher density, and the background area has a lower density, but it may contain a lot of false positive information. Therefore, median filtering can be used to remove spike noise in the image and highlight the edges and details of the image.

It should be noted that, in this embodiment of the present invention, after performing inverse perspective transformation on the ROI map in the road to be detected through S101 to obtain the transformed ROI map, the RGB color space model can also be directly used to perform the transformation on the transformed ROI map. The color probability map of lane line after processing is processed to obtain the processed color probability map of lane line, and then the clustering algorithm is used to perform regional segmentation processing on the color probability model map of lane line, and the result map of binary segmentation is obtained.

S105 , using a clustering algorithm to perform regional segmentation processing on the grayscale probability model graph of the lane line to obtain a binary segmentation result graph.

Illustratively, the clustering algorithm may be a partial information fuzzy C-means clustering algorithm, or a K-means clustering algorithm, and of course, other clustering algorithms may also be used. Illustratively, in this embodiment of the present invention, the clustering algorithm is a partial information fuzzy C-means clustering algorithm. The fuzzy C-means clustering algorithm in local information is a clustering algorithm that uses the membership degree of points to determine the degree of each point belonging to a certain cluster.

The specific algorithm is as follows:

计算第K类的中心；(1) First according to

Calculate the center of class K;

确定最小化目标函数；其中G_ki是模糊因数，

d_ij为像素i和j之间的空间欧氏距离；(2) According to

Determine the minimization objective function; where G _ki is the ambiguity factor,

d _ij is the spatial Euclidean distance between pixels i and j;

更新灰度值x_j相对于第k类的模糊隶属度，迭代上面步骤直到max|V^(b)-V^(b+1)|＜ε，其中V^(b)为模糊划分矩阵，b为循环次数，从而得到车道线的二值分割结果图。请参见图5所示，图5为本发明实施例提供的一种二值分割结果图的示意图。(3) Then according to

Update the fuzzy membership of the gray value x _j relative to the k-th class, and iterate the above steps until max|V ^(b) -V ^(b+1) |<ε, where V ^(b) is the fuzzy partition matrix, b is the cycle times to obtain the binary segmentation result map of the lane lines. Please refer to FIG. 5 , which is a schematic diagram of a binary segmentation result graph according to an embodiment of the present invention.

S106: Process the binary segmentation result graph, and acquire lane lines in the road to be detected.

After the binary segmentation result graph is obtained through S105, the binary segmentation result graph can be processed, and lane lines in the road to be detected are acquired. It can be seen that when obtaining the lane lines in the road to be detected, the YC _b C _r color space model is used to determine the transformed region of interest before using the clustering algorithm to segment the lane line gray probability model map. In the figure, the color probability of each pixel belonging to the lane line is normalized, and the gray probability of each pixel belonging to the lane line is obtained by normalizing the color probability of each pixel belonging to the lane line. The grayscale probability map of the lane line is obtained by probability, and then the generated grayscale probability map of the lane line is divided into regions, thereby improving the accuracy of the lane line detection.

An embodiment of the present invention provides a method for acquiring lane lines. First, inverse perspective transformation is performed on the ROI map on the road to be detected to obtain the transformed ROI map; the YC _b C _r color space model is used to determine the transformed ROI map. The color probability of each pixel belonging to the lane line in the area of interest map; and the color probability of each pixel belonging to the lane line is normalized to obtain the grayscale probability of each pixel belonging to the lane line; according to the fact that each pixel belongs to the lane The gray probability of the line is used to obtain the gray probability map of the lane line; the clustering algorithm is used to segment the gray probability model map of the lane line to obtain a binary segmentation result map; Lane lines in the road, thereby improving the accuracy of lane line detection.

Based on the embodiment shown in FIG. 1 , after the binary segmentation result graph is obtained, the binary segmentation result graph is processed, and the lane lines in the road to be detected are obtained. FIG. 6 is a schematic flowchart of processing a binary segmentation result graph according to an embodiment of the present invention.

S601. Perform morphological processing on the binary segmentation result graph to obtain a processed binary segmentation result graph.

Illustratively, in the embodiment of the present invention, by performing morphological processing on the binary segmentation result graph, the purpose is to cancel the peaks whose spans are smaller than the rectangular structure and remove noise, so that the lanes in the binary segmentation result graph after processing can be eliminated. The line area is more complete. Please refer to FIG. 7 , which is a schematic diagram of a binary segmentation result graph after morphological processing according to an embodiment of the present invention.

S602. Use the Sobel algorithm to detect the edge curve in the processed binary segmentation result graph.

Optionally, in the embodiment of the present invention, S602 adopts the Sobel algorithm to detect the edge curve in the processed binary segmentation result graph, which may include:

The Sobel algorithm is used to obtain the horizontal value and vertical value of each pixel in the processed binary segmentation result image; the gradient value and direction of each pixel are determined according to the horizontal value and vertical value of each pixel; when the gradient value of each pixel and When the direction is greater than the preset threshold, and the gradient value is within a certain preset range, determine that each pixel is an edge point of the processed binary segmentation result map; edge curve.

Specifically, the Sobel algorithm can accurately locate the edge in the low-contrast image, and the image after the approximate derivation of the horizontal and vertical directions is:

The gradient magnitude and direction are:

When the gradient of a pixel is greater than a certain threshold and the direction is within a certain range, the pixel is set as the edge point of the lane line area; after each edge point is determined, the processed binary value can be detected according to each edge point The edge curve in the segmentation result graph. Referring to FIG. 7 , FIG. 7 is a schematic diagram of an edge curve in a binary segmentation result graph after detection and processing by using the Sobel algorithm according to an embodiment of the present invention.

S603. Perform centerline operation processing on the edge curve to obtain a processed center curve.

The center line of the lane line area may be the average value of the two edge lines of the lane line area. Please refer to FIG. 8 , which is a schematic diagram of a processed center curve according to an embodiment of the present invention.

S604 , using Hall Hough transform to perform segmentation processing on the center curve, and performing fitting processing on a plurality of straight line segments obtained after the segmentation processing, to obtain lane lines in the road to be detected.

After obtaining the processed center curve through S603, the Hough transform can be used to search for straight lines in the left and right parts of the bottom area of the image respectively. Hough transform votes in the polar coordinate parameter space of the line, and the accumulated local peak points are candidate lines. Among them, the candidate straight line satisfies the following conditions:

(1) The angle between the straight line and the vertical direction must be within 25°; (2) The distance between the bottom intersection of the straight line and the bottom center point of the image cannot exceed the width of one lane in the image; (3) The two lane lines must be The parallel conditions are met; (4) When there are multiple straight lines that meet the conditions, select the straight line closest to the bottom center point on the left and right sides as the lane line.

According to the above four conditions, the most important straight line in the left and right parts of the bottom area of the image can be determined. In the current frame, the fixed-length line segment at the bottom of the detected line is selected as the initial segment of the lane lines on both sides; then the lane line is tracked along the direction of the initial segment; the entire lane line can be approximated as many straight line segments connected end to end; The angle of the lane line segment on the side can be used to approximately track the intermittent lane line on the other side; when the lane line is interrupted or there is a shadow on the road, the tracking program can also continuously track and detect the complete current frame lane line. Referring to FIG. 9 , FIG. 9 is a schematic diagram of lane lines on both sides obtained after tracking according to an embodiment of the present invention. After obtaining Figure 9, the least squares method can be used to fit a quadratic polynomial, and the fitted curve equation can be obtained by solving the coefficient matrix of the fitted polynomial. The fitting process converts the approximate tracking result into a high-precision fitting curve through a quadratic polynomial model, which improves the reliability of detection.

Optionally, in this embodiment of the present invention, after acquiring the lane lines in the road to be detected, it may further include:

Displays lane lines in the road to be detected on the map.

Using the Kalman filter to track the slope of the lane line for each frame, the post state can be predicted by the previous state and the current measurement, and the prediction is more accurate than the measurement alone, and then corrected frame by frame to draw an accurate global map, so as to display the pending state on the map. Detect lane lines in roads.

In order to verify the method for obtaining lane lines provided by the embodiment of the present invention, below, taking the relevant data in Table 1 as an example, select 300 frames in the video image sequence of the vehicle on the urban road for analysis, and the road quality in these images is relatively high. Poor, including different environments such as rain and night, but also a variety of different cornering situations. Among them, the size of each frame of image is 320*280, the average detection time of lane line per frame of image is 22 milliseconds, and the average is 45.5 frames per second. When the vehicle speed is 100km/h, that is, the vehicle travels 27.8m per second, the system updates the lane line every 0.61m. The test results are shown in Table 1.

Table 1 Detection accuracy in various environments

It can be seen from the detection results in Table 1 that there are 295 frames of images that can correctly detect lane lines, and the correct rate is 98.3%. When the lane lines are partially occluded, holes, and shadows, the system still has good detection results. Therefore, the lane line acquisition method provided by the embodiments of the present invention has good anti-interference performance.

FIG. 10 is a schematic structural diagram of an apparatus 100 for acquiring lane lines according to an embodiment of the present invention. Referring to FIG. 10 , the apparatus 100 for acquiring lane lines may include:

The transformation unit 1001 is configured to perform inverse perspective transformation on the ROI map on the road to be detected, to obtain the transformed ROI map.

The determining unit 1002 is configured to use the YC _b C _r color space model to determine the color probability that each pixel in the transformed region of interest map belongs to the lane line.

The processing unit 1003 is configured to perform normalization processing on the color probability that each pixel belongs to the lane line to obtain the grayscale probability that each pixel belongs to the lane line.

The generating unit 1004 is configured to obtain a grayscale probability map of the lane line according to the grayscale probability that each pixel belongs to the lane line.

The segmentation unit 1005 is configured to use a clustering algorithm to perform regional segmentation processing on the grayscale probability model graph of the lane line to obtain a binary segmentation result graph.

The obtaining unit 1006 is configured to process the binary segmentation result graph and obtain lane lines in the road to be detected.

确定变换后的感兴趣区域图中每一个像素属于车道线的颜色概率；Optionally, the determining unit 1002 is specifically configured to

Determine the color probability of each pixel belonging to the lane line in the transformed region of interest map;

为先验性概率，#C_i为颜色C_i类中的样本数，K表示第K种颜色。Among them, P(C _i |x) represents the probability that the pixel belongs to the lane line, N-1 represents that there are N-1 types of colors in the lane line, and C _i is the color of the ith type. The color of each pixel can be expressed as: x=( Y, C _b , C _r ), and x=(Y, C _b , C _r ), Y is the luminance component, C _b is the blue chrominance component, C _r is the red chrominance component, and P(x) is the pixel x The color probability of , P(x|C _i ) is the probability probability,

is the prior probability, #C _i is the number of samples in the color C _i class, and K represents the Kth color.

对每一个像素属于车道线的颜色概率进行归一化处理，得到每一个像素属于车道线的灰度概率；Optionally, the processing unit 1003 is specifically configured to

Normalize the color probability that each pixel belongs to the lane line, and obtain the gray probability that each pixel belongs to the lane line;

表示像素属于车道线的灰度概率。in,

Represents the grayscale probability that a pixel belongs to a lane line.

Optionally, the obtaining unit 1006 is specifically configured to perform morphological processing on the binary segmentation result graph to obtain the processed binary segmentation result graph; process the processed binary segmentation result graph, and obtain the road to be detected. Lane lines in .

Optionally, the obtaining unit 1006 is specifically configured to use the Sobel algorithm to detect the edge curve in the processed binary segmentation result graph; perform centerline operation processing on the edge curve to obtain the processed center curve; use Hall Hough The transformation performs segmentation processing on the center curve, and performs fitting processing on a plurality of straight line segments obtained after the segmentation processing, so as to obtain lane lines in the road to be detected.

Optionally, the obtaining unit 1006 is specifically used to obtain the horizontal value and vertical value of each pixel in the processed binary segmentation result graph by using the Sobel algorithm; determine the gradient value of each pixel according to the horizontal value and vertical value of each pixel. and direction; when the gradient value and direction of each pixel are greater than the preset threshold, and the gradient value is within a certain preset range, determine each pixel as the edge point of the processed binary segmentation result map; according to each edge point Detect edge curves in the processed binary segmentation result graph.

Optionally, the apparatus 100 for acquiring lane lines may further include a display unit 1007. Please refer to FIG. 11, which is a schematic structural diagram of another apparatus 100 for acquiring lane lines according to an embodiment of the present invention.

The display unit 1007 is used to display the lane lines in the road to be detected on the map.

The above-mentioned apparatus 100 for acquiring lane lines can correspondingly execute the technical solutions of the method for acquiring lane lines in any embodiment, and the implementation principles and technical effects thereof are similar, and details are not repeated here.

Other embodiments of the present disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common general knowledge or techniques in the technical field not disclosed by the present disclosure . The specification and examples are to be regarded as exemplary only, with the true scope and spirit of the disclosure being indicated by the following claims.

It is to be understood that the present disclosure is not limited to the precise structures described above and illustrated in the accompanying drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims (10)

1. A method for acquiring a lane line is characterized by comprising the following steps:

carrying out inverse perspective transformation on the interesting region image in the road to be detected to obtain a transformed interesting region image;

using YC_bC_rDetermining the color probability of each pixel in the transformed region of interest image belonging to the lane line by using a color space model;

normalizing the color probability of each pixel belonging to the lane line to obtain the gray level probability of each pixel belonging to the lane line;

obtaining a lane line gray probability map according to the gray probability of each pixel belonging to the lane line;

performing region segmentation processing on the lane line gray level probability model map by adopting a clustering algorithm to obtain a binary segmentation result map;

and processing the binary segmentation result graph, and acquiring a lane line in the road to be detected.

2. The method of claim 1, wherein the first and second light sources are selected from the group consisting of,characterized in that YC is adopted_bC_rThe color space model determines the color probability that each pixel in the transformed region of interest map belongs to the lane line, and the color space model comprises the following steps:

according to

Determining the color probability of each pixel in the transformed region of interest map belonging to a lane line;

wherein, P (C)_i| x) represents the probability that a pixel belongs to a lane line, N-1 represents that the lane line has N-1 types of colors in common, C_iFor the ith type of color, the color of each pixel can be expressed as: x ═ Y, C_b,C_r) And x ═ Y, C_b，C_r) Y is a luminance component, C_bIs the blue chrominance component, C_rIs the red chrominance component, P (x) is the color probability of pixel x, P (x | C)_i) In order to be the probability of a possibility,

for a priori probability, # C_iIs color C_iThe number of samples in a class, K represents the K-th color.

3. The method according to claim 2, wherein the normalizing the color probability of each pixel belonging to the lane line to obtain the gray level probability of each pixel belonging to the lane line comprises:

according to

Normalizing the color probability of each pixel belonging to the lane line to obtain the gray level probability of each pixel belonging to the lane line;

wherein,

representing the gray level probability of a pixel belonging to a lane line.

4. The method according to any one of claims 1 to 3, wherein the processing the binary segmentation result map and acquiring the lane lines in the road to be detected comprises:

performing morphological processing on the binary segmentation result graph to obtain a processed binary segmentation result graph;

and processing the processed binary segmentation result graph, and acquiring a lane line in the road to be detected.

5. The method according to claim 4, wherein the processing the processed binary segmentation result map and acquiring the lane lines in the road to be detected comprises:

detecting an edge curve in the processed binary segmentation result graph by adopting a Sobel algorithm;

performing centerline operation processing on the edge curves to obtain a processed image, wherein the processed image is provided with a center curve, and the center curve is the centerline of the two edge curves;

processing the processed image by adopting Hall Hough transformation according to the central curve to obtain initial sections of lane lines on two sides;

and tracking the lane lines along the initial section direction of the lane lines on the two sides according to the initial sections of the lane lines on the two sides, and fitting the lane lines on the two sides obtained after tracking to obtain the lane lines in the road to be detected.

6. The method according to claim 5, wherein the detecting the edge curve in the processed binary segmentation result map by using the Sobel algorithm comprises:

acquiring a horizontal value and a longitudinal value of each pixel in the processed binary segmentation result image by adopting the Sobel algorithm;

determining the gradient value and the direction of each pixel according to the transverse value and the longitudinal value of each pixel;

when the gradient value and the direction of each pixel are larger than a preset threshold value and the gradient value is within a certain preset range, determining each pixel as an edge point of the processed binary segmentation result image;

and detecting an edge curve in the processed binary segmentation result graph according to each edge point.

7. The method according to any one of claims 1 to 3, characterized in that after acquiring the lane line in the road to be detected, it comprises:

and displaying the lane lines in the road to be detected on a map.

8. An acquisition device of a lane line, comprising:

the transformation unit is used for carrying out inverse perspective transformation on the interesting region image in the road to be detected to obtain a transformed interesting region image;

a determination unit for adopting YC_bC_rDetermining the color probability of each pixel in the transformed region of interest image belonging to the lane line by using a color space model;

the processing unit is used for carrying out normalization processing on the color probability of each pixel belonging to the lane line to obtain the gray level probability of each pixel belonging to the lane line;

the generating unit is used for obtaining a lane line gray probability map according to the gray probability that each pixel belongs to the lane line;

the segmentation unit is used for carrying out region segmentation processing on the lane line gray level probability model map by adopting a clustering algorithm to obtain a binary segmentation result map;

and the acquisition unit is used for processing the binary segmentation result graph and acquiring the lane line in the road to be detected.

9. The apparatus of claim 8,

the determination unit is specifically used for determining

Determining the color probability of each pixel in the transformed region of interest map belonging to a lane line;

wherein, P (C)_i| x) represents the probability that a pixel belongs to a lane line, N-1 represents that the lane line has N-1 types of colors in common, C_iFor the ith type of color, the color of each pixel can be expressed as: x ═ Y, C_b,C_r) And x ═ Y, C_b，C_r) Y is a luminance component, C_bIs the blue chrominance component, C_rIs the red chrominance component, P (x) is the color probability of pixel x, P (x | C)_i) In order to be the probability of a possibility,

for a priori probability, # C_iIs color C_iThe number of samples in a class, K represents the K-th color.

10. The apparatus of claim 9,

the processing unit is specifically used for

Normalizing the color probability of each pixel belonging to the lane line to obtain the gray level probability of each pixel belonging to the lane line;

wherein,

representing the gray level probability of a pixel belonging to a lane line.

Images