Method, setup and device for determining lane lines

Abstract

Method for determining a lane line, comprising: determining lane line candidate areas in a road image by filtering the road image (201); extracting midpoints for each of the lane line candidate areas (202), wherein the midpoints include a midpoint between two endpoints of each line in each of the lane line candidate areas; obtaining a fitting curve of each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas (203); classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas (204);Dividing the lane line candidate areas into groups based on the classification result (205), wherein each curved lane line candidate area forms a group on its own, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line; and identifying two of the groups that satisfy a first predetermined condition as target groups, and identifying lane line candidate areas in the target groups as lane line areas (206), wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range.

Description

Area The present disclosure relates to the technical field of information processing and in particular to a method, a device and an apparatus for determining a lane line. background Document US 6,091,833 A describes a device and a method for accurately determining the position of a vehicle in a lane. A road contour signal and a road area signal are extracted from a digital image signal representing the view of the area in front of the vehicle. The lane position in the image signal is determined based on the road area signal and the road contour signal. The Wikipedia article "Lane Keeping Assist" (as of June 11, 2016, URL: https: / / de.wikipedia.org / w / index.php?title=Spurhalteassistent&oldid=155173247) describes various lane departure warning and lane keeping assist systems in motor vehicles. These systems use camera or infrared sensors to detect the lane and warn the driver when the vehicle is about to leave the lane, as well as active lane keeping assist systems with steering support.Document US 2015 / 0278613 A1 describes a lane marking detection device comprising an image acquisition device, a lane marking detection unit, and a lane marking search unit. The lane marking detection unit identifies a first and a second lane marking from image information. The search unit establishes search lines on the lane side of the second lane marking based on shape information from the first lane marking and determines brightness information from the image data along these search lines. Depending on whether a search line most likely matches a lane marking based on the brightness information is selected, this search line is chosen as a candidate lane marking, or no candidate line is selected. Many vehicles are currently equipped with a driver assistance system. Some driver assistance systems can provide lane information to vehicles by detecting lane markings on the road. For example, as an important part of an Advanced Driver Assistance System (ADAS), a lane departure warning system can detect lane markings, where a vehicle is traveling, and issue a warning to the driver if the vehicle deviates from the lane to ensure driving safety. In conventional technology, lane marking areas in a road image can be determined by filtering the image, and lane information can generally be recognized from the lane markings. However, according to the inventor's research, some of the lane marking areas obtained through filtering correspond to actual lane markings, while others do not. That is, not all lane marking areas determined using conventional technology are actually lane markings; rather, the determined lane marking areas also contain some interference areas that do not correspond to the actual lane markings. Therefore, the determination of lane marking areas using conventional technology is not accurate enough, leading to inaccurate determination of lane information and consequently preventing the lane departure warning system from issuing a precise warning message. Summary In light of the above, a method, a device and a apparatus for determining a lane line according to the present disclosure are created to filter out areas that do not correspond to real lane lines from lane line areas obtained by filtering, in order to make the determination of lane line areas and lane information more accurate, and thereby to enable a lane departure warning system to issue a warning message more accurately. In a first aspect, a procedure for determining a lane line is created, which includes: determining lane line candidate areas in a road image by filtering the road image; extracting center points for each of the lane line candidate areas, where the center points contain a midpoint between two endpoints of each line in each of the lane line candidate areas; obtaining a fitting curve of each of the lane line candidate areas by performing a curve fitting over the center points of each of the lane line candidate areas; classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas;Dividing the lane line candidate areas into groups based on the classification result, wherein each curved lane line candidate area forms a group, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line; and determining two of the groups that satisfy a first predetermined condition as target groups, and identifying lane line candidate areas in the target groups as lane line areas, wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. Optionally, two straight lane line candidate areas that satisfy a second predetermined condition are in the same group, wherein the second predetermined condition includes: that the matching line segments of the two straight lane line candidate areas do not overlap in a vertical direction, that the distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies within a second predetermined distance range, and that the directional deviation between the matching line segments of the two straight lane line candidate areas lies within a predetermined deviation range, wherein the matching line segments of the straight lane line candidate areas are obtained by performing straight-line matching over the centers of the straight lane line candidate areas. Optionally, after extracting midpoints for each of the lane line candidate areas, the procedure further includes: recognizing the number of effective pixels in the midpoints as an effective pixel count for each of the lane line candidate areas; and filtering out those of the lane line candidate areas that have smaller effective pixel counts than a predetermined count threshold, and for the remaining lane line candidate areas, performing the process of curve fitting over the midpoints of each of the lane line candidate areas. Optionally, obtaining the fit curve for each of the lane line candidate areas by performing a curve fit over the midpoints of each of the lane line candidate areas includes: selecting multiple control points from the midpoints of each of the lane line candidate areas, where the control points include midpoints at two ends of each of the lane line candidate areas; obtaining curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fit over the control points of each of the lane line candidate areas; and forming the fit curve for each of the lane line candidate areas using the curve equations of the pairs of adjacent control points of each of the lane line candidate areas. Optionally, selecting multiple control points from the midpoints of each of the lane line candidate areas includes: determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, selecting first target points from the lane line candidate area as the control points, where the distances between pairs of adjacent first target points in the lane line candidate area are equal and the first target points of the lane line candidate area have midpoints at two ends of the lane line candidate area;and in response to the existence of a reversal point at the midpoints of the lane line candidate area, selecting second target points from the lane line candidate area as the control points, wherein the second target points as the control points of the lane line candidate area include the midpoints at two ends of the lane line candidate area and the reversal point, the distances between any two adjacent second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and the distances between any two adjacent second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. Optionally, the procedure further includes: calculating a real length of each of the lane line candidate areas and determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; and identifying those of the lane line candidate areas whose real lengths are within a predetermined length range and whose midpoints do not contain a reversal point as straight lane line candidate areas, and for the remaining lane line candidate areas, performing the step of classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas. Optionally, classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each lane line candidate area includes: selecting a first sample point, a second sample point, and a third sample point from the centers of each lane line candidate area; calculating slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle, a second slope angle, and a third slope angle for each of the lane line candidate areas; and identifying, based on the first slope angle, the second slope angle, and the third slope angle of each of the lane line candidate areas, those of the lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition.Curved lane line candidate areas and other lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition, as straight lane line candidate areas, wherein the third predetermined condition is as follows: that both a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area are greater than a first gradient angle threshold, and wherein the fourth predetermined condition is as follows: that a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area is less than the first gradient angle threshold,and that a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area is greater than a second gradient angle threshold. Optionally, selecting the first, second, and third sample points from the centers of each of the lane line candidate areas includes: determining whether there is a reversal point at the centers of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the centers of any of the lane line candidate areas, selecting any three centers from the lane line candidate area as the first, second, and third sample points of the lane line candidate area;and in response to the fact that there is a reversal point in the midpoints of the lane line candidate area, determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and select two midpoints on two sides of the second sample point in the lane line candidate area as the first sample point and the third sample point of the lane line candidate area. Optionally, determining whether the reversal point exists at the midpoints of each of the lane line candidate areas includes: searching for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of each of the lane line candidate areas; in response to the third and fourth target points satisfying a fifth or sixth predetermined condition, determining that a reversal point exists at the midpoints of the corresponding lane line candidate area;and in response to the fact that the third and fourth destination points satisfy neither the fifth nor the sixth predetermined condition, determine that there is no reversal point at the midpoints of the corresponding lane line candidate area, wherein the fifth predetermined condition is as follows: that the third destination point is at the midpoint of the corresponding lane line candidate area, and the fourth destination point is at one of two ends of the corresponding lane line candidate area, and wherein the sixth predetermined condition is as follows: that the fourth destination point is at the midpoint of the corresponding lane line candidate area, and the third destination point is at one of two ends of the corresponding lane line candidate area. Optionally, the directional deviation between the matching line segments of the two straight lane line candidate areas, which lies within the predetermined deviation range, includes at least one of the following: that an inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range; that an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range; and that a deviation distance between the matching line segments of the two straight lane line candidate areas lies within a third predetermined distance range, where the deviation distance is a distance between a first intersection point and a second intersection point, the first intersection point being the first intersection point.The second intersection point is the intersection point of extension lines of the adjustment line segments of the two straight lane line candidate areas and a horizontal center line, and the distances between the nearest centers of the two straight lane line candidate areas in the vertical direction and the horizontal center line are equal. Optionally, the second predetermined condition further includes: that in a case where the inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined inclusion angle range, the deviation distance between the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined distance range, wherein the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined inclusion angle range. Optionally, after identifying lane line candidate areas in the target groups, the procedure further includes: performing an adjustment to the lane line areas in the target groups in order to obtain lane lines in the road image, whereby the lane line areas in each of the target groups are fitted into a lane line. In a second aspect, a device for determining a lane line is provided, comprising: a filter unit designed to determine lane line candidate areas in a road image by filtering the road image; an extraction unit designed to extract midpoints for each of the lane line candidate areas, wherein the midpoints contain a midpoint between two endpoints of each line in each of the lane line candidate areas; a first fitting unit designed to obtain a fitting curve of each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas; a classification unit designed to classify the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas;a grouping unit designed to divide the lane line candidate areas into groups based on the classification result, wherein each curved lane line candidate area forms a group on its own, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line;and a first detection unit designed to identify two of the groups satisfying a first predetermined condition as target groups and to identify lane line candidate areas in the target groups as lane line areas, wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. Optionally, two straight lane line candidate areas that satisfy a second predetermined condition are in the same group, wherein the second predetermined condition includes: that the matching line segments of the two straight lane line candidate areas do not overlap in a vertical direction, that the distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies within a second predetermined distance range, and that the directional deviation between the matching line segments of the two straight lane line candidate areas lies within a predetermined deviation range, wherein the matching line segments of the straight lane line candidate areas are obtained by performing straight-line matching over the centers of the straight lane line candidate areas. Optionally, the device further includes: a second detection unit designed, after extracting the midpoints for each of the lane line candidate areas, to recognize the number of effective pixels in the midpoints as an effective pixel count for each of the lane line candidate areas; and a filtering unit designed to filter out those of the lane line candidate areas that have smaller effective pixel counts than a predetermined number threshold, and to perform a curve fitting operation over the midpoints of each of the lane line candidate areas on the remaining lane line candidate areas. Optionally, the first fitting unit includes: a first selection unit, designed to select multiple control points from the midpoints of each of the lane line candidate areas, wherein the control points include midpoints at two ends of each of the lane line candidate areas; a second fitting unit, designed to obtain curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fit over the control points of each of the lane line candidate areas; and a formation unit, designed to form the fitted curve of each of the lane line candidate areas using the curve equations of the pairs of adjacent control points of each of the lane line candidate areas. Optionally, the first selection unit includes: a determination unit designed to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, wherein the reversal point is a peak or a trough; a second selection unit designed, in response to the absence of a reversal point at the midpoints of any of the lane line candidate areas, to select first target points from the lane line candidate area as the control points, wherein the distances between pairs of adjacent first target points in the lane line candidate area are equal and the first target points of the lane line candidate area include midpoints at two ends of the lane line candidate area;and a third selection unit, designed, in response to the existence of a reversal point at the midpoints of the lane line candidate area, to select second target points from the lane line candidate area as the control points in a case where there is a reversal point in the lane line candidate area, wherein the second target points as the control points of the lane line candidate area include midpoints at two ends of the lane line candidate area and the reversal point, distances between any two adjacent second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and distances between any two adjacent second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. Optionally, the device further includes: a first computation unit designed to calculate a real length of each of the lane line candidate areas and to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, the reversal point being a peak or a trough; and a third detection unit designed to identify as straight lane line candidate areas those of the lane line candidate areas whose real lengths are within a predetermined length range and whose midpoints do not contain a reversal point, and for the remaining lane line candidate areas, to perform the step of classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas. Optionally, the classification unit includes: a fourth selection unit, designed to select a first sample point, a second sample point, and a third sample point from the centers of each of the lane line candidate areas, wherein a distance between the first sample point and the second sample point is equal to a distance between the second sample point and the third sample point; a second calculation unit, designed to calculate slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle, a second slope angle, and a third slope angle for each of the lane line candidate areas; and a fourth detection unit, designed to determine, based on the first slope angle, the second slope angle, and the third slope angle of each of the lane line candidate areas, those of the lane line candidate areas.to identify lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition as curved lane line candidate areas, and other lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition as straight lane line candidate areas, wherein the third predetermined condition is as follows: that both a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area are greater than a first gradient angle threshold, and wherein the fourth predetermined condition is as follows: that a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area is less than the first gradient angle threshold.and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area is greater than a second gradient angle threshold. Optionally, the fourth selection unit includes: a determination unit designed to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; a fifth selection unit designed, in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, to select any three equally spaced midpoints from the lane line candidate area as the first sample point, the second sample point, and the third sample point of the lane line candidate area;and a sixth selection unit, designed, in response to the existence of a reversal point in the midpoints of the lane line candidate area, to determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and to select midpoints on two sides of the second sample point in the lane line candidate area as the first sample point and the third sample point of the lane line candidate area, wherein the distance from the midpoint selected as the first sample point to the second sample point is equal to the distance from the midpoint selected as the third sample point to the second sample point. Optionally, the determination unit includes: a search unit designed to search for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of each of the lane line candidate areas; a first determination unit designed, in response to the third and fourth target points satisfying a fifth predetermined condition or a sixth predetermined condition, to determine that there is a reversal point at the midpoints of the corresponding lane line candidate area;and a second determining unit, designed, in response to the fact that the third and fourth destination points satisfy neither the fifth predetermined condition nor the sixth predetermined condition, to determine that there is no reversal point at the midpoints of the corresponding lane line candidate area, wherein the fifth predetermined condition is as follows: that the third destination point is at the midpoint of the corresponding lane line candidate area and the fourth destination point is at one of two ends of the corresponding lane line candidate area, and wherein the sixth predetermined condition is as follows: that the fourth destination point is at the midpoint of the corresponding lane line candidate area and the third destination point is at one of two ends of the corresponding lane line candidate area.; Alternatively, the directional deviation between the matching line segments of the two straight lane line candidate areas, which lies within the predetermined deviation range, contains at least one of the following in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group: An inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range, an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range, and a deviation distance between the matching line segments of the two straight lane line candidate areas lies within a third predetermined distance range.where the deviation distance is a distance between a first intersection point and a second intersection point, the first intersection point and the second intersection point are intersection points of extension lines of the adjustment line segments of the two straight lane line candidate areas and a horizontal center line, and distances between nearest midpoints of the two straight lane line candidate areas in the vertical direction and the horizontal center line are equal. Optionally, in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group, the second predetermined condition further includes: that in a case where the inclusion angle of the matching line segments of the two straight lane line candidate areas is within a fourth predetermined inclusion angle range, the deviation distance between the matching line segments of the two straight lane line candidate areas is within a fourth predetermined distance range, wherein the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined inclusion angle range. Optionally, the device further includes: a third fitting unit, designed to perform fitting across the lane line areas in the target groups, after the lane line candidate areas in the target groups have been identified as the lane line areas, in order to obtain lane lines in the road image, whereby the lane line areas in each of the target groups are fitted into a lane line. In a third aspect, a device for determining a lane line has been created, which includes a processor, a memory, a communication interface and a bus system. The bus system is designed to couple the hardware of the device together. The communication interface is designed to establish a communication link between the device and at least one other device. The memory is designed to store program instructions and data. The processor is designed to read the instructions and data stored in memory and perform the following operations: Determining lane line candidate areas in a road image by filtering the road image; Extracting midpoints for each of the lane line candidate areas, where the midpoints contain a midpoint between two endpoints of each row in each of the lane line candidate areas; Obtaining a fitting curve for each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas; Classifying the lane line candidate areas into straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas;Dividing the lane line candidate areas into groups based on the classification result, wherein each curved lane line candidate area forms a group, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line; and determining two of the groups that satisfy a first predetermined condition as target groups, and identifying lane line candidate areas in the target groups as lane line areas, wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. Optionally, two straight lane line candidate areas that satisfy a second predetermined condition are in the same group, wherein the second predetermined condition includes: that the matching line segments of the two straight lane line candidate areas do not overlap in a vertical direction, that the distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies within a second predetermined distance range, and that the directional deviation between the matching line segments of the two straight lane line candidate areas lies within a predetermined deviation range, wherein the matching line segments of the straight lane line candidate areas are obtained by performing straight-line matching over the centers of the straight lane line candidate areas. Optionally, after extracting midpoints for each of the lane line candidate areas, the processor can further perform the following operations: recognizing the number of effective pixels in the midpoints as an effective pixel count for each of the lane line candidate areas; and filtering out those of the lane line candidate areas with smaller effective pixel counts than a predetermined count threshold, and for the remaining lane line candidate areas, performing the operation of a curve fitting over the midpoints of each of the lane line candidate areas. Optionally, the processor can obtain the fitting curve of each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas by performing the following operations: selecting multiple control points from the midpoints of each of the lane line candidate areas, where the control points include midpoints at two ends of each of the lane line candidate areas; obtaining curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fitting over the control points of each of the lane line candidate areas; and forming the fitting curve of each of the lane line candidate areas using the curve equations of the pairs of adjacent control points of each of the lane line candidate areas. Optionally, the processor can select the multiple control points from the midpoints of each of the lane line candidate areas by performing the following operations: Determine if there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, select first target points from the lane line candidate area as the control points, where the distances between pairs of adjacent first target points in the lane line candidate area are equal and the first target points of the lane line candidate area have midpoints at two ends of the lane line candidate area;and in response to the existence of a reversal point at the midpoints of the lane line candidate area, selecting second target points from the lane line candidate area as the control points, wherein the second target points as the control points of the lane line candidate area include the midpoints at two ends of the lane line candidate area and the reversal point, the distances between any two adjacent second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and the distances between any two adjacent second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. Optionally, the processor can further perform the following operations: calculating a real length of each of the lane line candidate areas and determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; and recognizing those of the lane line candidate areas whose real lengths are within a predetermined length range and whose midpoints do not contain a reversal point as straight lane line candidate areas, and for the remaining lane line candidate areas, performing the step of classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas. Optionally, the processor can classify the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each lane line candidate area by performing the following operations: selecting a first sample point, a second sample point, and a third sample point from the centers of each of the lane line candidate areas, where a distance between the first sample point and the second sample point is equal to a distance between the second sample point and the third sample point; calculating slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle, a second slope angle, and a third slope angle for each of the lane line candidate areas; and recognizing, based on the first slope angle,of the second gradient angle and the third gradient angle of each of the lane line candidate areas, those lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition are classified as curved lane line candidate areas, and other lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition are classified as straight lane line candidate areas, where the third predetermined condition is as follows: that both a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area are greater than a first gradient angle threshold,and wherein the fourth predetermined condition is as follows: that a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate range is less than the first gradient angle threshold and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate range is greater than a second gradient angle threshold. Optionally, the processor can select the first sample point, the second sample point, and the third sample point from the midpoints of each of the lane line candidate areas by performing the following operations: Determine if there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, select any three equally spaced midpoints from the lane line candidate area as the first sample point, the second sample point, and the third sample point of the lane line candidate area;and in response to the fact that there is a reversal point in the midpoints of the lane line candidate area, determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and select two midpoints on two sides of the second sample point in the lane line candidate area as the first sample point and the third sample point of the lane line candidate area, wherein the distance from the midpoint selected as the first sample point to the second sample point is equal to the distance from the midpoint selected as the third sample point to the second sample point. Alternatively, the processor can determine whether the reversal point exists at the midpoints of each of the lane line candidate areas by performing the following operations: searching for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of each of the lane line candidate areas; in response to the third and fourth target points satisfying a fifth or sixth predetermined condition, determining that a reversal point exists at the midpoints of the corresponding lane line candidate area;and in response to the fact that the third and fourth destination points satisfy neither the fifth nor the sixth predetermined condition, determine that there is no reversal point at the midpoints of the corresponding lane line candidate area, wherein the fifth predetermined condition is that the third destination point is at the midpoint of the corresponding lane line candidate area and the fourth destination point is at one of two ends of the corresponding lane line candidate area, and wherein the sixth predetermined condition is that the fourth destination point is at the midpoint of the corresponding lane line candidate area and the third destination point is at one of two ends of the corresponding lane line candidate area. Alternatively, the directional deviation between the matching line segments of the two straight lane line candidate areas, which lies within the predetermined deviation range, in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group, contains at least one of the following: that an inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range; that an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range; and that a deviation distance between the matching line segments of the two straight lane line candidate areas lies within a third predetermined distance range.where the deviation distance is a distance between a first intersection point and a second intersection point, the first intersection point and the second intersection point are intersection points of extension lines of the adjustment line segments of the two straight lane line candidate areas and a horizontal center line, and distances between nearest midpoints of the two straight lane line candidate areas in the vertical direction and the horizontal center line are equal. Alternatively, in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group, the second predetermined condition may further include: that in a case where the inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined inclusion angle range, the deviation distance between the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined distance range, where the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined inclusion angle range. Optionally, after recognizing lane line candidate areas in the target groups, the processor can further perform the following operation as lane line areas: Performing a fitting to the lane line areas in the target groups to obtain lane lines in the road image, whereby the lane line areas in each of the target groups are fitted into a lane line. In the present disclosure, the road image is filtered to obtain filtered areas as the lane line candidate areas. These lane line candidate areas are classified as straight lane line candidate areas and curved lane line candidate areas based on the fit curves of the lane line candidate areas and divided into groups based on the classification result, with the lane line candidate areas in each group forming a lane line. Two of the groups that correspond to matching lane lines are determined as target groups. The lane line candidate areas in the target groups are identified as the lane line areas.According to the above, after the lane line candidate areas are obtained by filtering the road image, interference areas that do not correspond to real lane lines can be removed from the lane line candidate areas by grouping and matching them. This allows areas within the lane line candidate areas that correspond to real lane lines to be recognized as lane line areas, making the determination of lane line areas more accurate and thus improving the determination of lane information. This, in turn, enables a lane departure warning system to issue an accurate warning message. Brief description of the drawing The drawing to be used in the description of the embodiment is briefly described below so that technical solutions according to the embodiments of the present disclosure become clearer. Obviously, the drawing in the following descriptions represents only some embodiments of the invention. Other drawings can be obtained by those skilled in the art based on this drawing. Fig. 1 is a schematic block diagram of a system involved in an application scenario according to one embodiment of the invention. Fig. 2 is a flowchart of a method for determining a lane line according to one embodiment of the invention. Fig. 3 is a schematic diagram of an example of lane line candidate areas in a road layout according to one embodiment of the invention.Figure 4 is a schematic diagram of an example of line segments formed by the centers of a lane line candidate area according to one embodiment of the invention. Figure 5 is a schematic diagram of an example of line segments formed by the centers of a lane line candidate area according to another embodiment of the invention. Figure 6 is a schematic diagram of an example of line segments formed by the centers of a lane line candidate area according to yet another embodiment of the invention. Figure 7 is a schematic diagram of an example of grouping straight lane line candidate areas according to one embodiment of the invention. Figure 8 is a schematic diagram of an exemplary distribution of adaptation line segments of lane line candidate areas in a road layout according to one embodiment of the invention.Figure 9 is a schematic diagram of an exemplary image coordinate system according to an embodiment of the invention. Figure 10 is a schematic diagram of an exemplary world coordinate system according to an embodiment of the invention. Figure 11 is a schematic diagram of a road image in an exemplary application scenario according to an embodiment of the invention. Figure 12 is a schematic diagram of filtered areas obtained by filtering a road image in an exemplary application scenario according to an embodiment of the invention. Figure 13 is a schematic diagram of filtered areas obtained by filtering the effective pixel count of interference areas according to an embodiment of the invention. Figure 14 is a schematic diagram of the grouping of lane line candidate areas in an application scenario according to an embodiment of the invention.Figure 15 is a schematic diagram of identified lane line candidate areas in an application scenario according to an embodiment of the invention. Figure 16 is a schematic diagram of a road layout with highlighted lane lines in an application scenario according to an embodiment of the invention. Figure 17 is a schematic block diagram of a device for determining a lane line according to an embodiment of the invention; and Figure 18 is a schematic block diagram of a device for determining a lane line according to an embodiment of the invention. Detailed description According to the inventor's research, conventional techniques for determining lane markings in a road image involve filtering the image to identify the lane markings. These filtered areas are then defined as the lane markings. However, due to the complexity of road surfaces, filtered areas often include curbs, fences, and water stains. Consequently, these filtered areas may contain not only actual lane markings but also other objects. As demonstrated, conventional methods for determining lane markings are insufficiently precise, leading to inaccurate lane information and preventing the lane departure warning system from issuing accurate warnings. To solve the problem, in the embodiments of the present disclosure, the road image is filtered to obtain filtered areas as the lane line candidate areas. The lane line candidate areas are classified as straight lane line candidate areas and curved lane line candidate areas according to the lane line candidate area fitting curves and divided into groups based on the classification result, with the lane line candidate areas in each group forming a lane line. Two of the groups that correspond to matching lane lines are determined as target groups. The lane line candidate areas in the target groups are identified as the lane line areas.According to the above, after the lane line candidate areas are obtained by filtering the road image, interference areas that do not correspond to real lane lines can be removed from the lane line candidate areas by grouping and matching them. This allows areas within the lane line candidate areas that correspond to real lane lines to be recognized as lane line areas, making the determination of lane line areas more accurate and thus improving the determination of lane information. This, in turn, enables a lane departure warning system to issue an accurate warning message. For example, the embodiment according to the present disclosure can be applied to a system shown in Fig. 1. For example, the system can be a driver assistance system and is more specifically, for example, a lane departure warning system. The system includes a processing device 101, a display device 102, and a receiving device 103. In the system, the receiving device 103 can receive a road image and send the road image to the processing device 101. The processing device 101 can determine lane line candidate areas in the road image by filtering the road image. The processing device 101 can extract midpoints for each of the lane line candidate areas, wherein the midpoints are the midpoints between two endpoints of each line in each of the lane line candidate areas.The processing device 101 can then obtain a fitting curve for each of the lane line candidate areas by performing a curve fitting over the midpoints of each lane line candidate area. The processing device 101 can then classify the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve for each lane line candidate area. Based on the classification result, the processing device 101 can divide the lane line candidate areas into groups, with each curved lane line candidate area forming a group, one or more straight lane line candidate areas forming a group, and the lane line candidate areas in the same group corresponding to the same lane line.Finally, the processing device 101 can identify two of the groups that meet a first predetermined condition as target groups and recognize the lane line candidate areas within the target groups as lane line areas. The processing device 101 can further generate a road image with highlighted lane lines based on the lane line areas and send the road image to the display device 102 for display. The first predetermined condition may contain at least one of the following: that a distance between lane lines corresponding to the two groups is less than a first predetermined distance threshold, and that an inclusion angle of the lane lines corresponding to the two groups is less than a first predetermined inclusion angle threshold. Understandably, the above scenario is only an example scenario according to the embodiments of the invention, and the embodiments of the invention are not limited to the above scenario. Certain embodiments of a method, a device and a apparatus for determining a lane line according to the embodiment of the present disclosure are described in detail below in conjunction with the drawing. Reference is made to Fig. 2, which shows a flowchart of a method for determining a lane line according to one embodiment of the invention. In this embodiment, the method may, for example, include the following steps 201 to 206. In step 201, lane line candidate areas are determined in a road image by filtering the road image. In practical implementations, a road image can be converted into a grayscale image, filtered, and the resulting filtered areas identified as candidate lane line regions. Understandably, obtaining these filtered areas by filtering the grayscale image can be achieved using any method. For example, the road image can be filtered using several conventional techniques, which are not described redundantly here. In step 202, midpoints are extracted for each of the lane line candidate areas, with the midpoints containing a midpoint between two endpoints of each row in each of the lane line candidate areas. Understandably, the entire road image consists of several pixel rows. Each lane line candidate area is a region within the road image; that is, each lane line candidate area actually occupies several pixel rows in the road image. For a lane line candidate area, there is a left endpoint and a right endpoint in each pixel row occupied by the lane line candidate area, and a midpoint between the left endpoint and the right endpoint of each row is the center point of the lane line candidate area. For example, in the image example shown in Fig. 3, each square represents a pixel in the image, with an area within a dashed box being a lane line candidate area. The lane line candidate area contains pixels a through o. Pixels a, d, g, j, and m are in a row. In this row, a left endpoint is a, a right endpoint is m, and a midpoint between them is g. Pixels b, e, h, k, and n are in a row. In this row, a left endpoint is b, a right endpoint is n, and a midpoint between them is h. Pixels c, f, i, l, and o are in a row. In this row, a left endpoint is c, a right endpoint is o, and a midpoint between them is i. Thus, the midpoints of the lane line candidate area are g, h, and i. Understandably, a real lane line area generally occupies a considerable number of pixel rows; a phantom lane line area may occupy only a few pixel rows, and the number of midpoints can represent the number of rows occupied by a lane line candidate area. Therefore, after extracting the midpoints for each lane line candidate area, and before any subsequent processing, some lane line candidate areas with few midpoints can be filtered out to eliminate some noisy areas from the lane line candidate areas obtained by filtering.More specifically, in some implementations of the embodiment, the procedure after step 202 may, for example, further include: recognizing the number of effective pixels at the centers as an effective pixel count for each of the lane line candidate areas; and filtering out lane line candidate areas with effective pixel counts exceeding a predetermined number range, and performing step 203 on the remaining lane line candidate areas. The predetermined number range may be a count threshold, and in the event that the number of effective pixels in a lane line candidate area is less than the count threshold, the lane line candidate area may be filtered out. In step 203, a fitting curve for each of the lane line candidate areas is obtained by performing a curve fitting over the midpoints of each of the lane line candidate areas. Understandably, for a lane line candidate area, the curve fitting via its midpoints can be performed by any possible implementation to obtain a fitting curve of the lane line candidate area. For example, in a specific example of curve fitting, in order for a fitting curve to more closely approximate the shape of a lane line candidate area, some control points can be selected from the midpoints of the lane line candidate area, and the curve fitting can be performed based on the control points.More specifically, in some implementations of the embodiment, step 203 may, for example, include: selecting multiple control points from the midpoints of each of the lane line candidate areas, wherein the control points include midpoints at two ends of each of the lane line candidate areas; obtaining curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fit over the control points of each of the lane line candidate areas by cubic spline interpolation; and forming the fit curve of each of the lane line candidate areas using the curve equations of the pairs of adjacent control points of each of the lane line candidate areas.The two ends of each of the lane line candidate areas are two edges of each of the lane line candidate areas in a vertical direction, and the midpoints at the two ends of each of the lane line candidate areas are midpoints of the two edges of each of the lane line candidate areas in the vertical direction. In other words, a midpoint with a maximum vertical coordinate and a midpoint with a minimum vertical coordinate from all the midpoints of each of the lane line candidate areas are the midpoints at the two ends of each of the lane line candidate areas. For example, in the example shown in Fig. 4, the line segment between a and e can be considered as a line segment formed by the midpoints of a lane line candidate area.Assuming that the centers a, b, c, d and e in the lane line candidate area are chosen as the control points, then a curve equation of a line segment between a and e, a curve equation of a line segment between b and c, a curve equation of a line segment between c and d and a curve equation of a line segment between d and e can be obtained by performing curve fitting over the control points a, b, c, d and e, and a curve equation corresponding to the lane line candidate area can be formed by combining the curve equations of the four line segments together, that is, forming the fitting curve of the lane line candidate area, where the centers a and e are centers at two ends of the lane line candidate area. Understandably, the adjustment can be performed using the control points of the lane line candidate area with curves of various shapes. For example, the adjustment can be performed using the cubic equation according to Formula 1: Formula 1 represents the curve equation of the first line segment in the lane line candidate area. In the curve equation, v and u are the vertical and horizontal coordinates in an image coordinate system based on the road surface image, and al, bl, cl and dl are parameters of the curve equation. It should be noted that if there is a reversal point at the midpoint of a lane line candidate area, sections of the lane line candidate area on either side of the reversal point will have highly different shapes and should be described by different fitting curves. To ensure that the fitting curve can be adaptive regardless of whether there is a reversal point or not, in some implementations of the embodiment the control points can be selected in different ways, depending on whether there is a reversal point at the midpoint of the lane line candidate area.More precisely, selecting multiple control points from the midpoints of each of the lane line candidate areas might involve: determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of a lane line candidate area, selecting first target points from the lane line candidate area as the control points; and in response to the finding that there is a reversal point at the midpoints of the lane line candidate area, selecting second target points from the lane line candidate area as the control points. The distances between pairs of adjacent first target points in the lane line candidate area are equal, and the first target points of the lane line candidate area have midpoints at two ends of the lane line candidate area.The second target points, as the control points of the lane line candidate area, include midpoints at two ends of the lane line candidate area and the reversal point. Distances between any two adjacent second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and distances between any two adjacent second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. For example, in the example shown in Fig. 4, the line segment between a and e can be considered as a line segment formed by the midpoints of a lane line candidate area. Since there is no reversal point at the midpoints of the lane line candidate area, a, b, c, d, and e can be chosen as the control points from the midpoints of the lane line candidate area; that is, the first target points contain a, b, c, d, and e. The distances between a and b, b and c, c and d, and d and e are equal. In the example shown in Fig.In the example shown, the line segment between a' and e' can be considered as a line segment formed by the midpoints of a lane line candidate area. Since there is a reversal point c' at the midpoints of the lane line candidate area, a', b', c', d', and e' can be chosen as the control points from the midpoints of the lane line candidate area; that is, the first target points include a', b', c', d', and e'. Here, a' is a midpoint at a first end of the lane line candidate area, e' is a midpoint at a second end of the lane line candidate area, the distances between a' and b' and b' are equal, and the distances between c' and d' and d' are equal. Understandably, at the midpoints of a lane line candidate area, a reversal point is either a peak or a trough with respect to the horizontal coordinate. More precisely, in an image coordinate system based on the road image, both a horizontal coordinate peak and a horizontal coordinate trough at the midpoints of a lane line candidate area are reversal points of that lane line candidate area. For a lane line candidate area, the horizontal coordinates of midpoints on two sides of a horizontal coordinate peak are all less than the horizontal coordinate of the horizontal coordinate peak, and the horizontal coordinates of midpoints on two sides of a horizontal coordinate trough are all greater than the horizontal coordinate of the horizontal coordinate trough. For example, in the one shown in Fig.In the example shown in Fig. 5, the line segment between a' and e' can be considered as a line segment formed by the midpoints of a lane line candidate area. At the midpoints of the lane line candidate area, c' is a horizontal coordinate trough. As a further example, in the example shown in Fig. 6, the line segment between a" and e" can be considered as a line segment formed by the midpoints of a lane line candidate area. At the midpoints of the lane line candidate area, c" is a horizontal coordinate apex. In this embodiment, the fact that there is a reversal point at the midpoints of a lane line candidate area can be determined by various implementations.For example, in some implementations of the embodiment, the fact that there is a reversal point in a lane line candidate area can be determined as follows: searching for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of the lane line candidate area; in response to the third and fourth target points satisfying a fifth or sixth predetermined condition, determining that there is a reversal point at the midpoints of the lane line candidate area; and in response to the third and fourth target points satisfying neither the fifth nor the sixth predetermined condition, determining that there is no reversal point at the midpoints of the lane line candidate area.The fifth predetermined condition is as follows: that the third target point lies in the middle of the lane line candidate area, and that the fourth target point lies at one of two ends of the lane line candidate area. The sixth predetermined condition is as follows: that the fourth target point lies in the middle of the lane line candidate area, and that the third target point lies at one of two ends of the lane line candidate area. Understandably, the horizontal coordinate here is the horizontal coordinate in the image coordinate system described above. More precisely, the fifth predetermined condition can be expressed by formulas 2 and 3: The sixth predetermined condition can be expressed by formulas 4 and 5: N represents the number of center points in a lane line candidate area, NumThre represents a predetermined number threshold, l max represents a vertical sequential number of a third target point in the center points of the lane line candidate area, and undimin represents a vertical sequential number of a fourth target point in the center points of the lane line candidate area. For example, in the image coordinate system built on the road image, assuming that the center points of the lane line candidate area are a, b, c, d, and e, in order from smallest to largest vertical coordinate, the vertical sequential number of a is 1, the vertical sequential number of b is 2, the vertical sequential number of c is 3, the vertical sequential number of d is 4, and the vertical sequential number of e is 5.Suppose a is the third target point with a maximum horizontal coordinate, and c is the fourth target point with a minimum horizontal coordinate, then l max is the vertical running number of a, that is, 1, and l min is the vertical running number of c, that is, 3. For a lane line candidate area, if a third target point and a fourth target point satisfy both of formulas 2 and 3, then the third target point and the fourth target point satisfy the fifth predetermined condition, and there is a peak point at the midpoints of the lane line candidate area. If the third target point and the fourth target point do not satisfy one or both of formulas 2 and 3, then the third target point and the fourth target point do not satisfy the fifth predetermined condition, and there is no peak point at the midpoints of the lane line candidate area. Similarly, if the third target point and the fourth target point satisfy one or both of formulas 4 and 5, then the third target point and the fourth target point satisfy the sixth predetermined condition, and there is a trough point at the midpoints of the lane line candidate area.If the third and fourth target points fail to satisfy one or both of formulas 4 and 5, then the third and fourth target points fail to satisfy the sixth predetermined condition, and there is no trough point at the midpoints of the lane line candidate area. Furthermore, if there is a peak or trough point at the midpoints of the lane line candidate area, then the peak or trough point is a reversal point therein. If there is no peak or trough point at the midpoints of the lane line candidate area, then there is no reversal point in the lane line candidate area. Understandably, after the fit curve of each of the lane line candidate areas is obtained in step 203, step 204 can be performed to classify the lane line candidate areas as straight and curved. That is, the fit curves obtained in step 203 are used to classify the shapes of the lane line candidate areas. A lane line candidate area can be identified as either straight or curved based on its fit curve. Furthermore, in some implementations of the embodiment, before performing step 204, some straight lane line candidate areas can be identified, for example, based on the actual lengths of the lane line candidate areas and whether a turning point exists. In this way, the shapes of lane line candidate areas with small actual lengths and no turning point do not need to be determined using fitting curves. Thus, the number of lane line candidate areas whose shapes need to be determined using fitting curves is reduced, thereby reducing the consumption of computing resources.More precisely, according to the embodiment, the method may further include, for example: calculating a real length of each of the lane line candidate areas and determining whether there is a reversal point at the midpoint of each of the lane line candidate areas, wherein the reversal point is a peak or a trough; and recognizing the lane line candidate areas whose real lengths lie within a predetermined length range and whose midpoints do not contain a reversal point as straight lane line candidate areas, and performing step 204 with the remaining lane line candidate areas. The predetermined length range may be embodied by a length threshold, and in the event that a real length of a lane line candidate area is less than the length threshold, the real length lies within the predetermined length range. In step 204, the lane line candidate areas are classified as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas. Understandably, the shape of a lane line candidate area corresponds to the shape of a lane line that corresponds to the lane line candidate area. More precisely, if a lane line candidate area is straight, a lane line corresponding to the lane line candidate area can be considered a straight line, and if a lane line candidate area is curved, a lane line corresponding to the lane line candidate area can be considered a curve. For a lane line candidate area, the shape of the lane line candidate area can be determined according to the slope angles at various points of its fitting curve. For example, in some implementations of the embodiment, three sample points can be selected from the centers of the lane line candidate area, slope angles of the fitting curve at the three sample points are calculated, and the shape of the lane line candidate area can be determined according to the slope angles at the three sample points. More precisely, step 204 can, for example, include: selecting a first sample point, a second sample point, and a third sample point from the centers of each of the lane line candidate areas; calculating slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle;a second gradient angle and a third gradient angle for each of the lane line candidate areas; and, based on the first gradient angle, the second gradient angle, and the third gradient angle of each of the lane line candidate areas, identifying those of the lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition as curved lane line candidate areas, and other of the lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition.as straight lane line candidate areas. The third predetermined condition is as follows: that both a difference between the first and second gradient angles of a lane line candidate area and a difference between the second and third gradient angles of the lane line candidate area are greater than a first gradient angle threshold; and the fourth predetermined condition is as follows: that a difference between the first and second gradient angles of a lane line candidate area is less than the first gradient angle threshold; and that a difference between the second and third gradient angles of the lane line candidate area is greater than a second gradient angle threshold. The first gradient angle threshold is less than the second gradient angle threshold. Understandably, the first sample point,The second and third sample points are located in an order from the largest to the smallest vertical coordinate in the image coordinate system; that is, the first, second, and third sample points are located from near to far with respect to the current position of a vehicle. More precisely, the first inclination angle, the second inclination angle and the third inclination angle can be calculated using formula 6: where θ0, θ1, and θ2 represent the first, second, and third slope angles, respectively, and k0, k1, and k2 represent a first, second, and third slope, respectively. The first slope is the slope of a tangent to the fitting curve of a lane line candidate area at the first sample point, the second slope is the slope of a tangent to the fitting curve of the lane line candidate area at the second sample point, and the third slope is the slope of a tangent to the fitting curve of the lane line candidate area at the third sample point. Furthermore, the third predetermined condition can be expressed by formulas 7 and 8: The fourth predetermined condition can be expressed by formulas 9 and 10: θthre1 represents the first slope angle threshold, θthre2 represents the second slope angle threshold, and θthre1 is smaller than θthre2. More precisely, θthre1 can be, for example, 3°, and θthre2 can be, for example, 8°. For a lane line candidate area, if the lane line candidate area fulfills formula 7 or 8, it satisfies the third predetermined condition and can be determined as a curved lane line candidate area. Similarly, if the lane line candidate area fulfills formula 9 or 10, it fulfills the fourth predetermined condition and can be determined as a curved lane line candidate area. If the lane line candidate area fulfills none of formulas 7, 8, 9, and 10, it fulfills neither the third nor the fourth predetermined condition and can be determined as a straight lane line candidate area. It should be noted that if there is a reversal point at the midpoint of a lane line candidate area, sections of the lane line candidate area on either side of the reversal point will have highly different shapes. Therefore, the existence of a reversal point can affect the determination of the shape of the lane line candidate area. Thus, in some implementations of the embodiment, the sample points can be selected in different ways depending on whether there is a reversal point at the midpoint of the lane line candidate area. More precisely, selecting a first sample point, a second sample point, and a third sample point from the midpoint of each of the lane line candidate areas might, for example, involve: determining whether there is a reversal point at the midpoint of each of the lane line candidate areas, where the reversal point is a peak or a trough;In response to the finding that there is no reversal point at the midpoints of a lane line candidate area, select any three midpoints from the lane line candidate area as the first, second, and third sample points of the lane line candidate area; and in response to the finding that there is a reversal point at the midpoints of the lane line candidate area, determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and select two midpoints on opposite sides of the second sample point in the lane line candidate area as the first and third sample points of the lane line candidate area. The first and third sample points could, for example, be midpoints at opposite ends of the lane line candidate area. For example, in the example shown in Fig. 4, the line segment between a and e can be considered as a line segment formed by the midpoints of a lane line candidate area. Since there is no reversal point at the midpoints of the lane line candidate area, a, cd, and e can be chosen as the sample points from the midpoints of the lane line candidate area; that is, a is the first sample point, c is the second sample point, and e is the third sample point. In the example shown in Fig.In the example shown in Figure 5, the line segment between a' and e' can be considered as a line segment formed by the midpoints of a lane line candidate area, and since there is a reversal point c' in the midpoints of the lane line candidate area, a', c' and e' can be chosen as the sample points from the midpoints of the lane line candidate area, that is, a' is the first sample point, c' is the second sample point, and e' is the third sample point. Furthermore, reference can be made to detailed descriptions of step 203 in the embodiment for introducing a reversal point, and how to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, which are not described redundantly here. In step 205, the lane line candidate areas are divided into groups based on the classification result, with each curved lane line candidate area forming a group on its own, one or more straight lane line candidate areas forming a group, and lane line candidate areas in the same group corresponding to the same lane line. Understandably, the lane line candidate areas in the road layout can be divided into different groups according to their shapes, so that lane line candidate areas corresponding to different lane lines can be placed in different groups. In this way, each group corresponds to a possible lane line, lane line candidate areas in the same group correspond to the same possible lane line, and lane line candidate areas in different groups do not correspond to the same lane line. In practical implementations, if there are curved lane line candidate areas in the road layout, each curved lane line candidate area can be placed in its own group, so that each group contains only a single curved lane line candidate area. For straight lane line candidate areas in the road layout, the straight lane line candidate areas can be divided into groups according to a second predetermined condition. More precisely, two straight lane line candidate areas that meet the second predetermined condition are in the same group. The second predetermined condition can be a condition to restrict a positional relationship and a directional deviation between two straight lane line candidate areas.More precisely, the second predetermined condition may include: that the fitting line segments of the two straight lane line candidate areas do not overlap in a vertical direction; that the distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies within a second predetermined distance range; and that the directional deviation between the fitting line segments of the two straight lane line candidate areas lies within a predetermined deviation range, wherein the fitting line segments of the straight lane line candidate areas are obtained by performing straight-line fitting over the centers of the straight lane line candidate areas.For example, a fitting line segment of a straight lane line candidate area can be obtained by performing a fitting calculation over midpoints of the straight lane line candidate area using the RANSAC least squares algorithm (matching a random sample). Understandably, in the second predetermined condition, the restriction of the positional relationship between the two straight lane line candidate areas can be described by a world coordinate system. More precisely, the fact that the adjustment line segments of the two straight lane line candidate areas do not overlap vertically can mean that areas of vertical coordinates of the two straight lane line candidate areas do not overlap in the world coordinate system constructed on the basis of the real world.The fact that the distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies in the second predetermined distance range can mean that a distance between two endpoints of the two straight lane line candidate areas that are closest to each other lies in the world coordinate system built on the basis of the real world. More precisely, the restriction of the positional relationship according to the second predetermined condition can be expressed by formulas 11 and 12: and Area 1 and Area 0 represent the two straight lane line candidate areas involved in the second predetermined condition, zS1 represents a vertical coordinate of a start point of the adjustment line segment of Area 1 in the world coordinate system, zE0 represents a vertical coordinate of an end point of the adjustment line segment of Area 0 in the world coordinate system, S represents a distance between the start point of the adjustment line segment of Area 1 and the end point of the adjustment line segment of Area 0 in the world coordinate system, MINBLANK represents the minimum distance threshold, MAXBLANK represents the maximum distance threshold, and MINBLANK and MAXBLANK may be determined according to a road lane design standard. Understandably, Area 1 is a region of the two straight lane line candidate regions far from the current position of the vehicle, and Area 0 is a region of the two straight lane line candidate regions near the current position of the vehicle. For an adaptation line segment, a starting point of this segment is an endpoint near the current position of the vehicle derived from two endpoints of the adaptation line segment, and an endpoint of this segment is an endpoint far from the current position of the vehicle derived from the two endpoints of the adaptation line segment. That is, the starting point of the adaptation line segment for Area 1 and the endpoint of the adaptation line segment for Area 0 are, in fact, two midpoints of Areas 1 and 0 that are closest to each other in the vertical direction. For example, in the exemplary grouping shown in Fig. 7, line segment “0” can serve as the adjustment line segment of region 0, and line segment “1” can serve as the adjustment line segment of region 1. E0 represents the endpoint of the adjustment line segment of region 0, S1 represents the starting point of region 1, E0 and S1 are the two nearest midpoints of the two regions in the vertical direction, a distance between E0 and S1 can serve as S in Formula 12, and vertical coordinates of E0 and S1 in the world coordinate system can serve as zE0 and zS1 in Formula 11. It should be noted that in the second predetermined condition, the restriction of the directional deviation of the two straight lane line candidate areas can be described in many ways.For example, in some embodiments, the requirement that a directional deviation between the matching line segments of the two straight lane line candidate areas lies within a predetermined deviation range may include any one or more of the following: Condition a: that an inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range; Condition b: that an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range; and Condition c: that a deviation distance between the matching line segments of the two straight lane line candidate areas lies within a third predetermined distance range. Condition a can be expressed by formula 13: Condition b can be expressed by formula 14: And condition c can be expressed by formula 15: Areas 1 and 0 represent the two straight lane line candidate areas involved in the second predetermined condition, α represents an inclusion angle of the matching line segment of area 1 and the matching line segment of area 0 in the image coordinate system, k1 represents a slope of the matching line segment of area 1, k0 represents a slope of the matching line segment of area 0, Threα represents an inclusion angle threshold to describe the second predetermined inclusion angle range, β represents an inclusion angle of the matching line segment of area 1 and the matching line segment of area 0 in the world coordinate system (that is, it is an estimated real inclusion angle), S is the same as S according to Formula 12, R represents a predetermined minimum rotation radius, Threβ represents an inclusion angle threshold to describe the third predetermined inclusion angle range.diff represents a deviation distance between the fitting line segment of range 1 and the fitting line segment of range 0, and Thrediff represents a distance threshold for describing the third predetermined distance range. For example, R can be 125 m, Thrediff can be 0.25 m, Threα can be 30°, and Threβ can be determined based on β, for example according to the equation, Understandably, the deviation distance is the distance between a first intersection point and a second intersection point. The first and second intersection points are the points where extension lines of the adjustment line segments of the two straight lane line candidate areas intersect a horizontal center line. The distances of the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction, to the horizontal center line, are equal. For example, in the exemplary grouping shown in Fig. 7, line segment “0” can serve as the adjustment line segment of area 0, and line segment “1” can serve as the adjustment line segment of area 1. E0 represents the endpoint of the adjustment line segment of area 0, S1 represents the starting point of area 1, and E0 and S1 are the two nearest midpoints of the two areas in the vertical direction. A distance between E0 and a dotted line “V=Vmiddle” is equal to a distance between S1 and the dotted line; that is, the dotted line “V=Vmiddle” is a horizontal centerline corresponding to the two areas. Point A is an intersection of an extension line of the adjustment line segment of area 1 and the horizontal centerline; that is, the first intersection point.Point B is the intersection of an extension line of the adjustment line segment of area 0 and the horizontal center line; that is, the second intersection point. Therefore, the distance between points A and B is the deviation distance between adjustment line segments of areas 1 and 0. There are sometimes some areas that do not correspond to any lane line in a road layout, and the adaptation line segments of these areas are approximately parallel to those of the areas that do correspond to lane lines, meaning they have similar gradients. For example, in the exemplary adaptation line segments of the lane line candidate areas shown in Fig. 8, line segments '1' and '2' belong to lane line areas that correspond to the same lane line, line segment '3' is an adaptation line segment of an area that corresponds to a fence, and line segment '3' lies almost parallel to segments '1' and '2', which have very similar gradients. To prevent an area that does not correspond to a lane line area and an area that corresponds to a lane line area from being grouped in the same group, in some implementations of the embodiment the second predetermined condition may further include, for example, condition d. The condition d is as follows: that in a case where the inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined inclusion angle range, the deviation distance between the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined distance range, wherein the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined inclusion angle range. More precisely, the condition can be expressed by formula 16: Reference can be made to the descriptions of formulas 13 to 15 for α and diff. Threα0 represents an inclusion angle threshold for describing the fourth predetermined inclusion angle range, and Thrediff0 represents a distance threshold for describing the fourth predetermined distance range. For the second predetermined inclusion angle range to be wider than the fourth predetermined inclusion angle range, Threα0 can be smaller than Threα, and for the third predetermined distance range to be wider than the fourth predetermined distance range, Thrediff0 can be smaller than Thrediff. It should be noted that for any two straight lane line candidate areas in the road image, if both straight lane line candidate areas satisfy the second predetermined condition, then both straight lane line candidate areas correspond to the same possible lane line and can be grouped together. For example, in the exemplary grouping shown in Fig. 7, each continuous line segment represents an adaptation line segment of a straight lane line candidate area. If line segments “0” and “1” satisfy the second predetermined condition, lane line candidate areas corresponding to line segments “0” and “1” are grouped together.If line segments "1" and "2" do not meet the second predetermined condition, lane line candidate areas corresponding to line segments "1" and "2" are divided into different groups. If line segments "1" and "3" meet the second predetermined condition, lane line candidate areas corresponding to line segments "1" and "3" are grouped into the same group.Understandably, if line segments "0" and "1" meet the second predetermined condition, and line segments "1" and "3" also meet the second predetermined condition, even though line segments "0" and "3" do not meet the second predetermined condition, line segments "0" and "3" are indeed grouped into the same group. This is because the lane line candidate areas corresponding to line segments "0" and "1" are grouped into the same group, and the lane line candidate areas corresponding to line segments "1" and "3" are also grouped into the same group. That is, line segments "0", "1", and "3" are indeed grouped into the same group. Understandably, there are often several straight lane line candidate areas in a road layout. These straight lane line candidate areas can be grouped in a specific order. For example, in some embodiments, all straight lane line candidate areas in the road layout can be arranged and numbered consecutively such that the starting points of adaptation line segments of the straight lane line candidate areas are located from near to far with respect to the current position of the vehicle. Then, in order from lowest to highest consecutive number, two straight lane line candidate areas are selected at a time to determine whether they satisfy the second predetermined condition. If both straight lane line candidate areas satisfy the second predetermined condition,The two straight lane line candidate areas are assigned to the same group. More precisely, step 205 can be performed, for example, by the following steps according to the embodiment: Step A: For all straight lane line candidate areas in the road image that have not been grouped, arrange the straight lane line candidate areas and assign consecutive numbers to the straight lane line candidate areas in an order such that the starting points of the adaptation line segments of the straight lane line candidate areas are located from near to far with respect to the current position of the vehicle; Step B: Set i=0 and j=i+1, create a new group as a current group, and add a straight lane line candidate area with a consecutive number i to the current group; Step C: Determine,Whether two straight lane line candidate areas with the serial numbers i and j satisfy the second predetermined condition; proceed to step D if the second predetermined condition is satisfied, and proceed to step G if the second predetermined condition is not satisfied; more precisely, determine whether formulas 11 to 16 are satisfied for the two straight lane line candidate areas, where it is determined that the two straight lane line candidate areas satisfy the second predetermined condition in a case where formulas 11 to 16 are satisfied; Step D: Determine the straight lane line candidate area with a serial number j as one area to be confirmed; Step E: Determine whether j equals N; proceed to step F if j does not equal N; proceed to step I if j equals N.where N represents the total number of straight lane line candidate areas; Step F: Determine if two straight lane line candidate areas with serial numbers j+1 and j overlap vertically, proceed to step H if the two straight lane line candidate areas with serial numbers j+1 and j overlap vertically, and proceed to step I if the two straight lane line candidate areas with serial numbers j+1 and j do not overlap vertically; Step G: Determine if j equals N, proceed to step H if j does not equal N, proceed to step I if j equals N; Step H: Set i=i and j=j+1 and return to step C; Step I: if the number of areas to be confirmed is currently 0, proceed directly to step J; in a case,that the number of areas to be confirmed is currently 1, proceed to step J after determining the areas to be confirmed as target areas, inserting the areas to be confirmed into the current group, and deleting the currently processed areas to be confirmed; and in a case that the number of areas to be confirmed is currently greater than 1, proceed to step J after a comprehensive analysis of the center uniformity of the areas to be confirmed, the slope differences between edge points on two sides of the areas to be confirmed, and the saturation levels of filtered areas, selecting a target area from the areas to be confirmed, inserting the target area into the current group, and deleting the currently processed areas to be confirmed; Step J: Determine if j equals N, proceed to step K in a case that j is not equal to N, and proceed to step L in a case where j is not equal to N.that j equals N; Step K: Set i to a maximum running number of straight lane line candidate areas in the current group and of j = j + 1 and return to Step C; and Step L: Return to Step A; the above steps are repeated until all straight lane line candidate areas in the road image are divided into groups. In step 206, two of the groups that meet the first predetermined condition are identified as target groups, and lane line candidate areas within the target groups are identified as lane line areas. The first predetermined condition may include at least one of the following: that a distance between lane lines corresponding to the two groups lies within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups lies within a first predetermined inclusion angle range.Understandably, two lane lines of the same lane are generally parallel to each other. If the angle between lane lines corresponding to two groups lies within a predetermined range, it is stated that the lane lines corresponding to the two groups can be considered approximately parallel. Furthermore, since it is subject to road design standards, the width of a lane is generally restricted within a certain range; that is, the spacing between lane lines of the same lane is restricted within a certain range. Therefore, if the distance between the lane lines corresponding to the two groups lies within the first predetermined spacing range, it is stated that the lane lines corresponding to the two groups meet lane spacing requirements.Furthermore, if the lane lines corresponding to the two groups meet area requirements for inclusion angle and spacing, the lane lines corresponding to the two groups can be considered as real lane lines, and lane line candidate areas in the two groups are real lane line areas. In some implementations of the embodiment, after determining the target groups and lane line areas in step 205, an adjustment can be made across the lane line areas in the target groups to obtain a lane line in the road image, where lane line areas in each target group are fitted into a single lane line. Understandably, the adjustment of the lane line areas can be performed, for example, using a cubic spline interpolation algorithm. Understandably, the image coordinate system constructed based on the road image according to the embodiment is a planar coordinate system within a planar road image, and one of its origins can be located, for example, at the upper left corner of the road image, as shown in the coordinate system in Fig. 9. The world coordinate system constructed based on the real world can be a three-dimensional coordinate system in a three-dimensional space where the vehicle is located, and one of its origins can be located, for example, at the position where a vertically downward-pointing line, starting from the center of the camera, intersects the ground, as shown in the coordinate system in Fig. 10. To help experts better understand the image processing workflow according to the embodiment, the process is described using an application scenario as an example. In this scenario, the road image is shown in Fig. 11, and an image of filtered areas obtained by filtering in step 201 is shown in Fig. 12. An image of filtered areas obtained by filtering the effective pixel count of noise areas is shown in Fig. 13, and the groups of lane line candidate areas obtained by splitting in step 205 are shown in Fig. 14, where areas with the same gray level belong to the same group. The lane line areas obtained by matching in step 206 are shown in Fig. 15, and lane lines obtained by fitting over the lane line areas are shown in Fig. 16. In the method according to the embodiment, the road image is filtered to obtain filtered areas as the lane line candidate areas. These lane line candidate areas are classified as straight lane line candidate areas and curved lane line candidate areas based on the fit curves of the lane line candidate areas and divided into groups based on the classification result, with the lane line candidate areas in each group forming a lane line. Two of the groups that correspond to matching lane lines are determined as target groups. The lane line candidate areas in the target groups are then identified as the lane line areas.According to the above, after the lane line candidate areas are obtained by filtering the road image, interference areas that do not correspond to real lane lines can be removed from the lane line candidate areas by grouping and matching them. This allows areas within the lane line candidate areas that correspond to real lane lines to be recognized as lane line areas, making the determination of lane line areas more accurate and thus improving the determination of lane information. This, in turn, enables a lane departure warning system to issue an accurate warning message. Reference is made to Fig. 17, which shows a schematic block diagram of a device for determining a lane line according to an embodiment of the invention. In this embodiment, the device may, for example, include: a filter unit 1701, designed to determine lane line candidate areas in a road image by filtering the road image; an extraction unit 1702, designed to extract center points for each of the lane line candidate areas, wherein the center points contain a midpoint between two endpoints of each row in each of the lane line candidate areas; a first fitting unit 1703, designed to obtain a fitting curve for each of the lane line candidate areas by performing a curve fitting over the center points of each of the lane line candidate areas;a classification unit 1704, designed to classify the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas; a grouping unit 1705, designed to divide the lane line candidate areas into groups based on the classification result, wherein each curved lane line candidate area alone forms a group, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line;and a first recognition unit 1706, designed to identify two of the groups satisfying a first predetermined condition as target groups and to identify lane line candidate areas in the target groups as lane line areas, wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. Optionally, two straight lane line candidate areas that satisfy a second predetermined condition are in the same group, wherein the second predetermined condition includes: that the matching line segments of the two straight lane line candidate areas do not overlap in a vertical direction, that the distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies within a second predetermined distance range, and that the directional deviation between the matching line segments of the two straight lane line candidate areas lies within a predetermined deviation range, wherein the matching line segments of the straight lane line candidate areas are obtained by performing straight-line matching over the centers of the straight lane line candidate areas. Optionally, the device further includes: a second detection unit designed, after extracting the midpoints for each of the lane line candidate areas, to recognize the number of effective pixels in the midpoints as an effective pixel count for each of the lane line candidate areas; and a filtering unit designed to filter out those of the lane line candidate areas that have smaller effective pixel counts than a predetermined number threshold, and to perform a curve fitting operation over the midpoints of each of the lane line candidate areas on the remaining lane line candidate areas. Optionally, the first fitting unit 1073 includes: a first selection unit designed to select multiple control points from the midpoints of each of the lane line candidate areas, wherein the control points include midpoints at two ends of each of the lane line candidate areas; a second fitting unit designed to obtain curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fit over the control points of each of the lane line candidate areas; and a formation unit designed to form the fitting curve of each of the lane line candidate areas using the curve equations of the pairs of adjacent control points of each of the lane line candidate areas. Optionally, the first selection unit includes: a determination unit designed to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, wherein the reversal point is a peak or a trough; a second selection unit designed, in response to the absence of a reversal point at the midpoints of any of the lane line candidate areas, to select first target points from the lane line candidate area as the control points, wherein the distances between pairs of adjacent first target points in the lane line candidate area are equal and the first target points of the lane line candidate area include midpoints at two ends of the lane line candidate area;and a third selection unit, designed, in response to the existence of a reversal point at the midpoints of the lane line candidate area, to select second target points from the lane line candidate area as the control points, wherein the second target points as the control points of the lane line candidate area include the midpoints at two ends of the lane line candidate area and the reversal point, distances between any two adjacent second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and distances between any two adjacent second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. Optionally, the device further includes: a first computation unit designed to calculate a real length of each of the lane line candidate areas and to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, the reversal point being a peak or a trough; and a third detection unit designed to identify as straight lane line candidate areas those of the lane line candidate areas whose real lengths are within a predetermined length range and whose midpoints do not contain a reversal point, and for the remaining lane line candidate areas, to perform the step of classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas. Optionally, the classification unit 1704 includes: a fourth selection unit, designed to select a first sample point, a second sample point, and a third sample point from the centers of each of the lane line candidate areas, wherein a distance between the first sample point and the second sample point is equal to a distance between the second sample point and the third sample point; a second calculation unit, designed to calculate slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle, a second slope angle, and a third slope angle for each of the lane line candidate areas; and a fourth detection unit, designed to identify, based on the first slope angle, the second slope angle, and the third slope angle of each of the lane line candidate areas, those of the lane line candidate areas.to identify lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition as curved lane line candidate areas and other lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition as straight lane line candidate areas, wherein the third predetermined condition is as follows: that both a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area are greater than a first gradient angle threshold, and wherein the fourth predetermined condition is as follows: that a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area is less than the first gradient angle threshold.and that a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area is greater than a second gradient angle threshold. Optionally, the fourth selection unit includes: a determination unit designed to determine whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; a fifth selection unit designed, in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, to select any three equally spaced midpoints from the lane line candidate area as the first sample point, the second sample point, and the third sample point of the lane line candidate area;and a sixth selection unit, designed, in response to the existence of a reversal point in the midpoints of the lane line candidate area, to determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and to select two midpoints on two sides of the second sample point in the lane line candidate area as the first sample point and the third sample point of the lane line candidate area, wherein the distance from the midpoint selected as the first sample point to the second sample point is equal to the distance from the midpoint selected as the third sample point to the second sample point. Optionally, the determination unit includes: a search unit designed to search for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of each of the lane line candidate areas; a first determination unit designed, in response to the third and fourth target points satisfying a fifth predetermined condition or a sixth predetermined condition, to determine that there is a reversal point at the midpoints of the corresponding lane line candidate area;and a second determining unit, designed, in response to the fact that the third destination point and the fourth destination point satisfy neither the fifth predetermined condition nor the sixth predetermined condition, to determine that there is no reversal point at the midpoints of the corresponding lane line candidate area, wherein the fifth predetermined condition is as follows: that the third destination point lies in the middle of the corresponding lane line candidate area, and that the fourth destination point lies at one of two ends of the corresponding lane line candidate area, and wherein the sixth predetermined condition is as follows: that the fourth destination point lies in the middle of the corresponding lane line candidate area, and that the third destination point lies at one of the two ends of the corresponding lane line candidate area.; Alternatively, the directional deviation between the matching line segments of the two straight lane line candidate areas, which lies within the predetermined deviation range, contains at least one of the following in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group: An inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range, an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range, and a deviation distance between the matching line segments of the two straight lane line candidate areas lies within a third predetermined distance range.where the deviation distance is a distance between a first intersection point and a second intersection point, the first intersection point and the second intersection point are intersection points of extension lines of the adjustment line segments of the two straight lane line candidate areas and a horizontal center line, and distances between nearest midpoints of the two straight lane line candidate areas in the vertical direction and the horizontal center line are equal. Optionally, in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group, the second predetermined condition further includes: that in a case where the inclusion angle of the matching line segments of the two straight lane line candidate areas is within a fourth predetermined inclusion angle range, the deviation distance between the matching line segments of the two straight lane line candidate areas is within a fourth predetermined distance range, wherein the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined distance range. Optionally, the device further includes: a third fitting unit, designed to perform a fitting across the lane line areas in the target groups, after the lane line candidate areas in the target groups have been identified as the lane line areas, in order to obtain lane lines in the road image, whereby the lane line areas in each of the target groups are fitted into a lane line. With the device according to the embodiment, after the lane line candidate areas have been obtained by filtering the road image, interference areas that do not correspond to real lane lines can be removed from the lane line candidate areas by grouping and matching the lane line candidate areas, so that areas in the lane line candidate areas that correspond to real lane lines can be recognized as lane line areas, which makes the determination of lane line areas more accurate and thus makes the determination of lane information more accurate and thereby enables a lane departure warning system to issue a warning message accurately. Reference is made to Fig. 18, which shows a schematic block diagram of a device for determining a lane line according to an embodiment of the invention. In this embodiment, the device 1800 can, for example, include a processor 1801, a memory 1802, a communication interface 1803, and a bus system 1804. The 1804 bus system is designed to couple the hardware of the device together. The communication interface 1803 is designed to maintain a communication link between the device and at least one other device. The 1802 memory is designed to store program instructions and data. The 1801 processor is designed to read the instructions and data stored in memory and perform the following operations: Determining lane line candidate areas in a road image by filtering the road image; Extracting midpoints for each of the lane line candidate areas, where the midpoints contain a midpoint between two endpoints of each row in each of the lane line candidate areas; Obtaining a fitting curve for each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas; Classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas;Dividing the lane line candidate areas into groups based on the classification result, wherein each curved lane line candidate area forms a group, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line; and determining two of the groups that satisfy a first predetermined condition as target groups, and identifying lane line candidate areas in the target groups as lane line areas, wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. Optionally, two straight lane line candidate areas that satisfy a second predetermined condition are in the same group, wherein the second predetermined condition includes: that the matching line segments of the two straight lane line candidate areas do not overlap in a vertical direction, that a distance between the centers of the two straight lane line candidate areas that are closest to each other in the vertical direction lies within a second predetermined distance range, and that a directional deviation between the matching line segments of the two straight lane line candidate areas lies within a predetermined deviation range, wherein the matching line segments of the straight lane line candidate areas are obtained by performing straight-line matching over the centers of the straight lane line candidate areas. Optionally, after extracting the midpoints for each of the lane line candidate areas, the 1801 processor can further perform the following operations: recognizing the number of effective pixels in the midpoints as an effective pixel count for each of the lane line candidate areas; and filtering out those of the lane line candidate areas that have smaller effective pixel counts than a predetermined count threshold, and for the remaining lane line candidate areas, performing the operation of a curve fitting over the midpoints of each of the lane line candidate areas. Alternatively, the 1801 processor can obtain the fit curve of each of the lane line candidate areas by performing a curve fit over the midpoints of each of the lane line candidate areas by performing the following operations: selecting multiple control points from the midpoints of each of the lane line candidate areas, where the control points include midpoints at two ends of each of the lane line candidate areas; obtaining curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fit over the control points of each of the lane line candidate areas; and forming the fit curve of each of the lane line candidate areas using the curve equations of the pairs of adjacent control points of each of the lane line candidate areas. Optionally, the 1801 processor can select the multiple control points from the midpoints of each of the lane line candidate areas by performing the following operations: Determine if there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, select first target points from the lane line candidate area as the control points, where the distances between pairs of adjacent first target points in the lane line candidate area are equal and the first target points of the lane line candidate area include midpoints at two ends of the lane line candidate area;and in response to the existence of a reversal point at the midpoints of the lane line candidate area, selecting second target points from the lane line candidate area as the control points, wherein the second target points as the control points of the lane line candidate area include midpoints at two ends of the lane line candidate area and the reversal point, the distances between any two adjacent second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and the distances between any two adjacent second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. Optionally, the 1801 processor can further perform the following operations: calculating a real length of each of the lane line candidate areas and determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; and recognizing those of the lane line candidate areas whose real lengths are within a predetermined length range and whose midpoints do not contain a reversal point as straight lane line candidate areas, and for the remaining lane line candidate areas, performing the step of classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas. Optionally, the processor can classify the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each lane line candidate area by performing the following operations: selecting a first sample point, a second sample point, and a third sample point from the centers of each of the lane line candidate areas, where a distance between the first sample point and the second sample point is equal to a distance between the second sample point and the third sample point; calculating slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle, a second slope angle, and a third slope angle for each of the lane line candidate areas; and recognizing, based on the first slope angle,of the second gradient angle and the third gradient angle of each of the lane line candidate areas, those lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition are classified as curved lane line candidate areas, and other lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition are classified as straight lane line candidate areas, where the third predetermined condition is as follows: that both a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area are greater than a first gradient angle threshold,and wherein the fourth predetermined condition is as follows: that a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate range is less than the first gradient angle threshold, and that a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate range is greater than a second gradient angle threshold. Alternatively, the 1801 processor can select the first sample point, the second sample point, and the third sample point by performing the following operations: Determine if there is a reversal point at the midpoints of each of the lane line candidate areas, where the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, select any three equally spaced midpoints from the lane line candidate area as the first sample point, the second sample point, and the third sample point of the lane line candidate area;and in response to the fact that there is a reversal point in the midpoints of the lane line candidate area, determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and select two midpoints on two sides of the second sample point in the lane line candidate area as the first sample point and the third sample point of the lane line candidate area, wherein the distance from the midpoint selected as the first sample point to the second sample point is equal to the distance from the midpoint selected as the third sample point to the second sample point. Alternatively, the 1801 processor can determine whether the reversal point exists at the midpoints of each of the lane line candidate areas by performing the following operations: searching for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of each of the lane line candidate areas; in response to the third and fourth target points satisfying a fifth or sixth predetermined condition, determining that a reversal point exists at the midpoints of the corresponding lane line candidate area;and in response to the fact that the third and fourth destination points satisfy neither the fifth nor the sixth predetermined condition, determine that there is no reversal point at the midpoints of the corresponding lane line candidate area, wherein the fifth predetermined condition is that the third destination point lies in the middle of the corresponding lane line candidate area, and that the fourth destination point lies at one of two ends of the corresponding lane line candidate area, and wherein the sixth predetermined condition is that the fourth destination point lies in the middle of the corresponding lane line candidate area, and that the third destination point lies at one of the two ends of the corresponding lane line candidate area. Alternatively, the directional deviation between the matching line segments of the two straight lane line candidate areas, which lies within the predetermined deviation range, in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group, contains at least one of the following: that an inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range; that an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range; and that a deviation distance between the matching line segments of the two straight lane line candidate areas lies within a third predetermined distance range.where the deviation distance is a distance between a first intersection point and a second intersection point, the first intersection point and the second intersection point are intersection points of extension lines of the adjustment line segments of the two straight lane line candidate areas and a horizontal center line, and distances between nearest midpoints of the two straight lane line candidate areas in the vertical direction and the horizontal center line are equal. Alternatively, in a case where the two straight lane line candidate areas that satisfy the second predetermined condition are in the same group, the second predetermined condition may further include: that in a case where the inclusion angle of the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined inclusion angle range, the deviation distance between the matching line segments of the two straight lane line candidate areas lies within a fourth predetermined distance range, where the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined distance range. Optionally, after recognizing the lane line candidate areas in the target groups, the processor 1801 can perform the following operation: performing a fitting on the lane line areas in the target groups to obtain lane lines in the road image, whereby the lane line areas in each of the target groups are fitted into a lane line. With the device according to the embodiment, after the lane line candidate areas have been obtained by filtering the road image, interference areas that do not correspond to real lane lines can be removed from the lane line candidate areas by grouping and matching the lane line candidate areas, so that areas in the lane line candidate areas that correspond to real lane lines can be recognized as lane line areas, which makes the determination of lane line areas more accurate and thus makes the determination of lane information more accurate and thereby enables a lane departure warning system to issue a warning message accurately. The terms “first(s)” in “first predetermined condition”, “first target point”, and “first predetermined distance range”, which are referenced in the embodiments according to the present disclosure, are used only as a descriptive term and do not represent the first in a sequence. The same applies to “second(s)” and “third(s)”. Based on the foregoing descriptions of the embodiments, those skilled in the art can clearly understand that all or some of the steps of the method according to the above embodiments can be performed by software plus a hardware platform. Based on this understanding, the technical solutions according to the present disclosure can be embodied in the form of a software product, and the computer software product can be stored on a storage medium, such as read-only memory (ROM), RAM, a disk, or a compact disc, containing a number of instructions for a computer device (which may be a personal computer, a server, or a network communication device, such as a router) to implement the methods according to the embodiments or some parts thereof. The embodiments according to the present disclosure are described progressively; reference may be made to one another with regard to identical or similar parts, and each embodiment highlights differences from the others. In particular, the method embodiments and the device embodiments are thus briefly described, since they are fundamentally similar to the system embodiments, and reference may be made to the descriptions of the system embodiments with regard to these parts.The foregoing device and system embodiments are for illustrative purposes only; modules described as separate components may or may not be physically separate, and components depicted as modules may or may not be physical modules, that is, they may be located in one place or distributed across several network units, with some or all of the modules being used as practically required to achieve the purpose of the technical solutions according to the embodiments. Persons skilled in the art can understand and implement the above without creative effort. The foregoing embodiments are only some preferred embodiments of the invention and are not intended to limit the scope of this disclosure. It should be emphasized that improvements and enhancements can be made by those skilled in the art without departing from the invention, and that these improvements and enhancements are intended to fall within the scope of this disclosure.

Claims

Method for determining a lane line, comprising: determining lane line candidate areas in a road image by filtering the road image (201); extracting midpoints for each of the lane line candidate areas (202), wherein the midpoints include a midpoint between two endpoints of each line in each of the lane line candidate areas; obtaining a fitting curve of each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas (203); classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas (204);Dividing the lane line candidate areas into groups based on the classification result (205), wherein each curved lane line candidate area forms a group on its own, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line; and identifying two of the groups that satisfy a first predetermined condition as target groups, and identifying lane line candidate areas in the target groups as lane line areas (206), wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. The method of claim 1, wherein two straight lane line candidate areas satisfying a second predetermined condition are in the same group, the second predetermined condition comprising: that adjustment line segments of the two straight lane line candidate areas do not overlap each other in a vertical direction, that a distance between centers nearest to each other in the vertical direction of the two straight lane line candidate areas lies within a second predetermined distance range, and that a directional deviation between the adjustment line segments of the two straight lane line candidate areas lies within a predetermined deviation range, and wherein the adjustment line segments of the straight lane line candidate areas are obtained by performing straight-line adjustments over centers of the straight lane line candidate areas. The method of claim 1, wherein the method further comprises, after extracting center points for each of the lane line candidate areas (202): recognizing the number of effective pixels in the center points as an effective pixel count for each of the lane line candidate areas; and filtering out such lane line candidate areas that have smaller effective pixel counts than a predetermined number threshold, and for the remaining lane line candidate areas, performing the process of curve fitting via the center points of each of the lane line candidate areas. The method of claim 1, wherein obtaining the fit curve of each of the lane line candidate areas by performing a curve fit over the midpoints of each of the lane line candidate areas comprises: selecting a plurality of control points from the midpoints of each of the lane line candidate areas, wherein the control points comprise midpoints at two ends of each of the lane line candidate areas; obtaining curve equations of pairs of adjacent control points in each of the lane line candidate areas by performing a fit over the control points of each of the lane line candidate areas; and forming the fit curve of each of the lane line candidate areas with the curve equations of the pairs of adjacent control points of each of the lane line candidate areas. The method of claim 4, wherein selecting the plurality of control points from the midpoints of each of the lane line candidate areas comprises: determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, wherein the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, selecting first target points from the lane line candidate area as the control points, wherein the distances between the pairs of adjacent first target points in the lane line candidate area are equal and the first target points of the lane line candidate area comprise midpoints at two ends of the lane line candidate area;and in response to the existence of a reversal point at the midpoints of the lane line candidate area, select second target points from the lane line candidate area as the control points, wherein the second target points as the control points of the lane line candidate area include the midpoints at two ends of the lane line candidate area and the reversal point, the distances between adjacent two second target points located between a midpoint at a first end of the lane line candidate area and the reversal point are equal, and the distances between adjacent two second target points located between a midpoint at a second end of the lane line candidate area and the reversal point are equal. The method of claim 1, further comprising: calculating a real length of each of the lane line candidate areas and determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, wherein the reversal point is a peak or a trough; and identifying such lane line candidate areas whose real lengths are within a predetermined length range and whose midpoints do not contain a reversal point as straight lane line candidate areas, and for the remainder of the lane line candidate areas, performing step (204) of classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas. The method of claim 1, wherein classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to the slope angle of the fitting curve of each of the lane line candidate areas (204) comprises: selecting a first sample point, a second sample point, and a third sample point from the centers of each of the lane line candidate areas; calculating slope angles of the fitting curve at the first sample point, at the second sample point, and at the third sample point as a first slope angle, a second slope angle, and a third slope angle for each of the lane line candidate areas; and identifying, based on the first slope angle, the second slope angle, and the third slope angle of each of the lane line candidate areas, such of the lane line candidate areas thatThose lane line candidate areas that satisfy a third predetermined condition or a fourth predetermined condition are classified as curved lane line candidate areas, and other lane line candidate areas that satisfy neither the third predetermined condition nor the fourth predetermined condition are classified as straight lane line candidate areas, wherein the third predetermined condition is that both a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area and a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area are greater than a first gradient angle threshold, and wherein the fourth predetermined condition is that a difference between the first gradient angle and the second gradient angle of the corresponding lane line candidate area is less than the first gradient angle threshold.and that a difference between the second gradient angle and the third gradient angle of the corresponding lane line candidate area is greater than a second gradient angle threshold. The method of claim 7, wherein selecting the first sample point, the second sample point, and the third sample point from the midpoints of each of the lane line candidate areas comprises: determining whether there is a reversal point at the midpoints of each of the lane line candidate areas, wherein the reversal point is a peak or a trough; in response to the finding that there is no reversal point at the midpoints of any of the lane line candidate areas, selecting any three midpoints from the lane line candidate area as the first sample point, the second sample point, and the third sample point of the lane line candidate area;and in response to the fact that there is a reversal point in the midpoints of the lane line candidate area, determine the reversal point in the lane line candidate area as the second sample point of the lane line candidate area and select two midpoints on two sides of the second sample point in the lane line candidate area as the first sample point and the third sample point of the lane line candidate area. The method of claim 5, 6 or 8, wherein determining whether the reversal point exists at the midpoints of each of the lane line candidate areas comprises: searching for a third target point with a maximum horizontal coordinate and a fourth target point with a minimum horizontal coordinate at the midpoints of each of the lane line candidate areas; in response to the fact that the third target point and the fourth target point satisfy a fifth predetermined condition or a sixth predetermined condition, determining that there is a reversal point at the midpoints of the corresponding lane line candidate area;and in response to the fact that the third and fourth destination points satisfy neither the fifth nor the sixth predetermined condition, determine that there is no reversal point at the midpoints of the corresponding lane line candidate area, wherein the fifth predetermined condition is that the third destination point lies in the middle of the corresponding lane line candidate area, and that the fourth destination point lies at one of two ends of the corresponding lane line candidate area, and wherein the sixth predetermined condition is that the fourth destination point lies in the middle of the corresponding lane line candidate area, and that the third destination point lies at one of the two ends of the corresponding lane line candidate area. The method of claim 2, wherein the directional deviation between the adaptation line segments of the two straight lane line candidate areas, which lies within the predetermined deviation range, comprises at least one of the following: that an inclusion angle of the adaptation line segments of the two straight lane line candidate areas lies within a second predetermined inclusion angle range, that an estimated actual inclusion angle of the two straight lane line candidate areas lies within a third predetermined inclusion angle range, and that a deviation distance between the adaptation line segments of the two straight lane line candidate areas lies within a third predetermined distance range, wherein the deviation distance is a distance between a first intersection point and a second intersection point, the first intersection point being the first intersection point and the second intersection point being the second intersection point.The second intersection point is the intersection point of extension lines of the adjustment line segments of the two straight lane line candidate areas and a horizontal center line, and the distances between the nearest centers of the two straight lane line candidate areas in the vertical direction and the horizontal center line are equal. The method of claim 2, wherein the second predetermined condition further comprises: that in a case where the inclusion angle of the adaptation line segments of the two straight lane line candidate areas lies within a fourth predetermined inclusion angle range, the deviation distance between the adaptation line segments of the two straight lane line candidate areas lies within a fourth predetermined distance range, wherein the second predetermined inclusion angle range is wider than the fourth predetermined inclusion angle range and the third predetermined distance range is wider than the fourth predetermined distance range. The method according to claim 1, wherein the method after recognizing lane line candidate areas in the target groups further comprises: performing a fitting on the lane line areas in the target groups to obtain lane lines in the road image, wherein the lane line areas in each of the target groups are fitted into a lane line. A device for determining a lane line, comprising: a filter unit (1701) designed to determine lane line candidate areas in a road image by filtering the road image; an extraction unit (1702) designed to extract midpoints for each of the lane line candidate areas, wherein the midpoints comprise a midpoint between two endpoints of each line in each of the lane line candidate areas; a first fitting unit (1703) designed to obtain a fitting curve of each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas; a classification unit (1704) designed to classify the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas;a grouping unit (1705) designed to divide the lane line candidate areas into groups based on the classification result, wherein each curved lane line candidate area alone forms a group, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line;and a first recognition unit (1706), designed to identify two of the groups satisfying a first predetermined condition as target groups and to identify lane line candidate areas in the target groups as lane line areas, wherein the first predetermined condition comprises at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range. Device (1800) for determining a lane line, comprising a processor (1801), a memory (1802), a communication interface (1803) and a bus system (1804), wherein the bus system (1804) is designed to couple hardware of the device (1800); the communication interface (1803) is designed to establish a communication link between the device (1800) and at least one other device; the memory (1802) is designed to store program instructions and data; and the processor (1801) is designed to read the instructions and data stored in the memory (1802) and to perform the following operations: determining lane line candidate areas in a road image by filtering the road image (201);Extracting midpoints for each of the lane line candidate areas (202), wherein the midpoints include a midpoint between two endpoints of each line in each of the lane line candidate areas; Obtaining a fitting curve of each of the lane line candidate areas by performing a curve fitting over the midpoints of each of the lane line candidate areas (203); Classifying the lane line candidate areas as straight lane line candidate areas and curved lane line candidate areas according to a slope angle of the fitting curve of each of the lane line candidate areas (204);Dividing the lane line candidate areas into groups based on the classification result (205), wherein each curved lane line candidate area forms a group on its own, one or more straight lane line candidate areas form a group, and lane line candidate areas in the same group correspond to the same lane line; and identifying two of the groups that satisfy a first predetermined condition as target groups, and identifying lane line candidate areas in the target groups as lane line areas (206), wherein the first predetermined condition includes at least one of the following: that a distance between lane lines corresponding to the two groups is within a first predetermined distance range, and that an inclusion angle of the lane lines corresponding to the two groups is within a first predetermined inclusion angle range.

Images