System and method for extracting point cloud road lane
The point cloud road lane extraction system addresses the challenge of lane recognition for autonomous vehicles by analyzing point cloud data to ensure lane continuity and safe driving, even with dotted lines, through data filtering and continuity detection, enhancing the accuracy and reliability of lane detection.
Patent Information
- Application Number
- PCT/KR2024/005999
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-19
- Filing Date
- 2024-05-03
- Publication Date
- 2024-11-28
AI Technical Summary
Autonomous vehicles face challenges in recognizing and maintaining lane continuity, especially with dotted lines, which can lead to deviation from the lane and potential accidents due to limitations in existing lane recognition technologies, particularly under insufficient lighting conditions.
A point cloud road lane extraction system and method that analyzes point cloud data using a combination of data filtering, lane extraction, and continuity detection units to create a continuous detection object, enabling the identification of lane continuity and interpolation for seamless lane detection, even when lanes are blurred or missing.
The system ensures safe autonomous driving by accurately extracting connected linear shapes of lanes, maintaining lane continuity, and completing alignments, thereby preventing deviations from designated lanes.
Smart Images

Figure KR2024005999_28112024_PF_FP_ABST
Abstract
Description
Point cloud road lane extraction system and method
[0001] The present invention relates to a point cloud road lane extraction system and method, and more particularly, to a point cloud road lane extraction system and method that analyzes point cloud (PCL) data collected on a road to extract lanes, creates continuous detection objects on the lanes, and determines continuity between the extracted lanes, thereby extracting a connected linear shape for a dotted line lane, thereby enabling an autonomous vehicle to drive safely without departing from the dotted line lane.
[0002]
[0003] Autonomous vehicles are vehicles capable of sensing their surroundings and movements with little or no human intervention or input. To realize these autonomous vehicles, a variety of sensors capable of sensing their surroundings, such as radar, computer vision, GPS, odometry, and inertial measurement units, must be integrated.
[0004] Furthermore, technologies are needed to provide autonomous vehicles with vehicle operation support information, enabling them to easily understand surrounding traffic conditions that are difficult for autonomous vehicles to detect. For example, Local Dynamic Map (LDM) is a technology that links, stores, and manages standardized vehicle operation support information for autonomous cooperative driving. Based on a lane-level precision electronic map (or static information), it integrates road traffic and surrounding vehicle conditions (dynamic information) to provide real-time information to autonomous vehicles.
[0005] The unmanned autonomous driving of a vehicle (autonomous vehicle) can be broadly divided into three stages: recognizing the surrounding environment, planning a driving route based on the recognized environment, and driving along the planned route.
[0006] The technology for recognizing the environment around a vehicle will vary depending on the target environment in which the autonomous vehicle is driving. In particular, in order to drive autonomously in a road environment designed and constructed for existing manned vehicles, technology for recognizing various rules existing on the road may be essential.
[0007] In particular, recognizing lanes and following a designated lane is the most basic technology for safe driving with manned vehicles.
[0008] While there are methods for recognizing road lanes based on images captured by cameras, they suffer from reduced recognition accuracy when ambient light is insufficient. Consequently, lane recognition technology using Lidar is currently being developed.
[0009] Lidar is a device that precisely draws the surroundings by measuring the distance to objects and other things by emitting laser pulses and then reflecting them back. It has the same principle as radar using radio waves, but since it uses the frequency of the visible light range, which is called light among electromagnetic waves, the actual technology and scope of use can be seen to be different.
[0010] These lidars are one component of the Mobile Mapping System (MMS), and in addition to lidars, the MMS includes a camera, an inertial measurement unit (IMU), GPS, and a distance measurement indicator (DMI).
[0011] Accordingly, MMS can obtain precise digital map data by matching the information obtained through lidar with the coordinates obtained from GPS.
[0012] When using point cloud data collected via lidar to recognize lanes, the most commonly used lanes on roads, which are suitable for overtaking, are typically broken lines, leaving gaps in the middle of the lane. If an autonomous vehicle perceives these gaps as not being lanes, it could drift out of its lane, potentially leading to an accident. Therefore, for autonomous vehicles to drive along broken line lanes, it is crucial to detect the continuity between lanes.
[0013] To this end, lane information is acquired in advance through the construction of a high-definition map (HD map) and transmitted to the vehicle, and the vehicle goes through a verification step by comparing the information from the real-time lane recognition sensor with the lane information of the HD map constructed in advance.
[0014]
[0015] The purpose of the present invention is to provide a point cloud road lane extraction system and method that extracts lanes by analyzing point cloud (PCL) data collected on a road, creates continuous detection objects on the lanes, and determines continuity between the extracted lanes, thereby extracting a connected linear shape for a dotted line lane, thereby enabling an autonomous vehicle to drive safely without leaving the dotted line lane.
[0016] The purpose of the present invention is not limited to the purposes mentioned above, and other purposes not mentioned will be clearly understood by those skilled in the art from the description below.
[0017]
[0018] According to one embodiment of the present invention, a point cloud road lane extraction system comprises: a point cloud data storage unit that stores point cloud (PCL) data collected on a road; a lane extraction unit that analyzes the point cloud data to extract a lane; and a continuous detection unit that creates a continuous detection object in a first lane extracted by the lane extraction unit and detects whether the continuous detection object overlaps a second lane, thereby determining continuity between the first lane and the second lane.
[0019] The above continuous detection object may have a triangular shape with one vertex positioned within the first lane and a base opposite to the vertex facing in the direction of travel of the road.
[0020] The second lane is located in front of the first lane in the direction of travel of the road, and the continuous detection unit can determine continuity between the first lane and the second lane by detecting whether the base of the triangle of the continuous detection object overlaps the second lane.
[0021] The above point cloud road lane extraction system further includes a data filtering unit that filters point cloud data stored in the point cloud data storage unit, and the lane extraction unit can extract a lane by analyzing the point cloud data filtered through the data filtering unit.
[0022] The above data filtering unit can perform filtering by batch-shifting the signal intensity values of point cloud data stored in the point cloud data storage unit and amplifying and trimming the deviation between the signal intensity values.
[0023] The above data filtering unit can perform a drive near filter that extracts only point cloud data within a predetermined distance based on a driving line converted into a line by collecting points at points that are divided left and right when point cloud data stored in the point cloud data storage unit is collected on a road.
[0024] The above data filtering unit can perform a slice filter that groups point cloud data that has undergone the drive near filter by slicing it into groups at regular intervals in the direction of travel of the road.
[0025] The above point cloud road lane extraction system may further include a missing interpolation unit that provides continuity to missing lanes based on continuity between lanes including the first lane and the second lane determined by the continuous detection unit, thereby completing a connected line between lanes of the road.
[0026] The above lane extraction unit can extract lanes by using a K-means algorithm that calculates the proximity state between points of the point cloud data and groups adjacent points to divide traffic surface marking areas on the road, calculating the minimum bounding box of the divided traffic surface marking areas, and applying a linear extraction algorithm to divide linear components.
[0027] According to one embodiment of the present invention, a point cloud road lane extraction method comprises a point cloud data storage step of storing point cloud (PCL) data collected on a road; a lane extraction step of analyzing the point cloud data to extract a lane; and a continuous detection step of creating a continuous detection object in a first lane extracted in the lane extraction step and detecting whether the continuous detection object overlaps a second lane to determine continuity between the first lane and the second lane.
[0028] The above point cloud road lane extraction method may further include a data filtering step of filtering point cloud data stored in the point cloud data storage step, and the lane extraction step may further include a step of analyzing point cloud data filtered through the data filtering step to extract lanes.
[0029] The above data filtering step may include a step of performing filtering by batch-shifting the signal intensity values of the point cloud data stored in the point cloud data storing step and amplifying and trimming the deviation between the signal intensity values.
[0030] The above data filtering step may include a step of performing a drive near filter to extract only point cloud data within a predetermined distance based on a driving line converted into a line by collecting points at left and right points when the point cloud data stored in the point cloud data storing step is collected on the road.
[0031] The above data filtering step may include a step of performing a slice filter that groups point cloud data on which the drive near filter has been performed by slicing the data into groups at regular intervals in the direction of travel of the road.
[0032] The above point cloud road lane extraction method may further include a missing interpolation step for completing a connected line between lanes of the road by granting continuity to missing lanes based on continuity between lanes including the first lane and the second lane determined in the continuous detection step.
[0033]
[0034] According to one aspect of the present invention, a point cloud road lane extraction system and method can be provided, which extracts lanes by analyzing point cloud (PCL) data collected on a road, creates continuous detection objects on the lanes, and determines continuity between the extracted lanes, thereby extracting a connected linear shape for a dotted line lane, thereby enabling an autonomous vehicle to drive safely without leaving the dotted line lane.
[0035] And, the point cloud road lane extraction system and method according to one embodiment of the present invention performs missing interpolation to complete a connected line between road lanes by granting continuity to missing lanes based on continuity between lanes, thereby extracting a connected line even when the lane does not exist, is blurred, or is erased.
[0036] In addition, the point cloud road lane extraction system and method according to one embodiment of the present invention performs a slice filter that groups point cloud data by slicing it at regular intervals in the direction of travel of the road, thereby enabling seamless lane detection.
[0037] The effects that can be obtained from the present invention are not limited to the effects mentioned above, and other effects not mentioned can be clearly understood by a person having ordinary skill in the art to which the present invention belongs from the description below.
[0038]
[0039] FIG. 1 is a schematic diagram illustrating a point cloud road lane extraction system according to one embodiment of the present invention.
[0040] FIG. 2a and FIG. 2b are diagrams illustrating an example of point cloud data stored in a point cloud data storage unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0041] FIG. 3a and FIG. 3b are diagrams illustrating an example of point cloud data from which noise has been removed by a data filtering unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0042] FIG. 4a and FIG. 4b are diagrams illustrating an example of point cloud data to which a drive-near filter is applied by a data filtering unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0043] FIG. 5 is a diagram illustrating an example of point cloud data to which a slice filter is applied by a data filtering unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0044] FIG. 6a and FIG. 6b are diagrams illustrating an example of point cloud data grouped by a lane extraction unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0045] FIG. 7 is a drawing specifically illustrating a process of determining continuity by a continuous detection unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0046] FIG. 8a and FIG. 8b are drawings illustrating an example of a continuous detection unit of a point cloud road lane extraction system according to one embodiment of the present invention generating a continuous detection object to detect continuity between lanes.
[0047] FIG. 9a and FIG. 9b are diagrams illustrating an example of a result of detecting continuity between lanes by a continuous detection unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0048] FIG. 10a and FIG. 10b are diagrams illustrating an example of a result of a missing interpolation unit of a point cloud road lane extraction system according to one embodiment of the present invention completing a connected line between lanes.
[0049] Fig. 11 is a flowchart for explaining a point cloud road lane extraction method according to one embodiment of the present invention.
[0050] FIG. 12 is a block diagram illustrating a computer system for implementing a point cloud road lane extraction method according to one embodiment of the present invention.
[0051]
[0052] The advantages and features of the present invention, and the methods for achieving them, will become clear with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and these embodiments are provided only to make the disclosure of the present invention complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Meanwhile, the terminology used in this specification is for the purpose of describing the embodiments and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated in the phrase. The terms "comprises" and / or "comprising" as used in the specification do not exclude the presence or addition of one or more other components, steps, operations, and / or elements mentioned.
[0053] In describing the present invention, if it is determined that a detailed description of a related known technology may unnecessarily obscure the gist of the present invention, the detailed description is omitted.
[0054] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. In order to facilitate an overall understanding in describing the present invention, the same reference numbers will be used for the same means regardless of the drawing numbers.
[0055]
[0056] FIG. 1 is a schematic diagram illustrating a point cloud road lane extraction system according to one embodiment of the present invention.
[0057] Referring to FIG. 1, a point cloud road lane extraction system (100) according to one embodiment of the present invention is configured to include a point cloud data storage unit (110), a data filtering unit (120), a lane extraction unit (130), a continuous detection unit (140), and a missing interpolation unit (150). The point cloud road lane extraction system (100) illustrated in FIG. 1 is according to one embodiment, and the components of the point cloud road lane extraction system (100) according to the present invention are not limited to the embodiment illustrated in FIG. 1, and may be added, changed, or deleted as needed.
[0058]
[0059] The point cloud data storage unit (110) stores point cloud (PCL) data collected on the road. Point cloud data is data collected by Lidar sensors, RGB-D sensors, etc. These sensors send light / signals to objects and record the time it takes for them to return, calculate distance information for each light / signal, and create a single point. Point cloud data refers to a set cloud of multiple points spread across a three-dimensional space collected from these sensors. In one embodiment, the point cloud data stored in the point cloud data storage unit (110) may be data collected from a Lidar sensor equipped in a mobile mapping system (MMS). Hereinafter, an example of point cloud data stored in the point cloud data storage unit (110) will be described with reference to FIGS. 2A and 2B .
[0060]
[0061] FIG. 2a and FIG. 2b are diagrams illustrating an example of point cloud data stored in a point cloud data storage unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0062] Referring to FIGS. 2A and 2B , point cloud data that can be stored in the point cloud data storage unit (110) in the las file format as data collected by the lidar sensor of the MMS vehicle is illustrated. As illustrated in FIGS. 2A and 2B , point cloud data including information related to not only lanes or road markings on the road but also trees, structures, tunnels, etc. around the road are collected through the lidar sensor. In addition, due to the characteristics of the las file format, when point cloud data is collected on the road through the lidar sensor, the data is collected by being divided left and right along the driving path of the MMS vehicle. Therefore, when point cloud data is collected on the road, a driving line (210) that is converted into a line by gathering points at the time points divided left and right can be expressed on the point cloud data.
[0063]
[0064] Returning to FIG. 1, the data filtering unit (120) filters point cloud data stored in the point cloud data storage unit. The data filtering unit (120) is used to reduce the amount of data processed when extracting lanes and detecting continuity between lanes in the point cloud road lane extraction system (100) according to one embodiment of the present invention, thereby increasing processing speed and enhancing the accuracy of the processing results.
[0065] In one embodiment, the data filtering unit (120) can perform filtering by batch shifting the signal intensity values of the point cloud data stored in the point cloud data storage unit (110) and amplifying and trimming the deviation between the signal intensity values. The point cloud data (.las) collected from the MMS vehicle is composed of a single record of 3D point coordinates composed of x, y, and z, signal intensity, GPS time, etc. First, the data filtering unit (120) batch shifts the signal intensity values. This is the same concept as the brightness filter in image processing. Then, the data filtering unit (120) performs a process of amplifying and trimming the deviation between the signal intensity values. This is the same concept as the contrast filter in image processing. Through this process, the data filtering unit (120) can remove noise from the point cloud data. Referring to FIGS. 3a and 3b below, point cloud data from which noise has been removed will be described.
[0066]
[0067] FIG. 3a and FIG. 3b are diagrams illustrating an example of point cloud data from which noise has been removed by a data filtering unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0068] Referring to FIGS. 3a and 3b, it can be seen that, compared to FIGS. 2a and 2b, the point cloud data is brighter due to the signal intensity value being shifted through a brightness filter, and noise is removed through a contrast filter.
[0069]
[0070] Returning to FIG. 1 again, in one embodiment, the data filtering unit (120) may perform a drive near filter to extract only point cloud data within a predetermined distance based on a driving line (210), which is a line converted by collecting points at the left and right points when point cloud data stored in the point cloud data storage unit (110) is collected on the road. The point cloud data also includes information related to trees, structures, etc. around the road, which is unnecessary for extracting lanes. Therefore, the data filtering unit (120) may retain only the point cloud data necessary for extracting lanes through the drive near filter. In one embodiment, the predetermined distance for the data filtering unit (120) to extract point cloud data based on the driving line (210) may be a distance necessary for leaving only road-related information in the point cloud data, and may vary depending on the width of the road. Hereinafter, point cloud data to which the drive near filter is applied will be described with reference to FIGS. 4A and 4B.
[0071]
[0072] FIG. 4a and FIG. 4b are diagrams illustrating an example of point cloud data to which a drive-near filter is applied by a data filtering unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0073] Referring to FIGS. 4a and 4b, it can be confirmed that a drive near filter is applied by the data filtering unit (120) of the point cloud road lane extraction system (100) according to one embodiment of the present invention, thereby extracting only point cloud data within a predetermined distance based on the driving line (210), thereby removing information related to trees, tunnels, etc. around the road.
[0074]
[0075] Returning to FIG. 1, in one embodiment, the data filtering unit (120) may perform a slice filter that slices and groups point cloud data that has undergone a drive-near filter at regular intervals in the direction of travel of the road. The point cloud road lane extraction system (100) according to one embodiment of the present invention is for detecting continuity between lanes when the lanes are formed by dotted lines. The method of detecting continuity between lanes in the point cloud road lane extraction system (100) according to one embodiment of the present invention, which will be described in detail later, is performed on the premise that the intervals between lanes are constant. However, on actual roads, lanes are a mixture of dotted and solid lines, and the intervals between lanes formed by dotted lines are not constant. Therefore, when the method of detecting continuity between lanes in the point cloud road lane extraction system (100) according to one embodiment of the present invention is applied to an actual road, a situation may occur where continuity detection is interrupted in a section where a solid line appears in the middle or the intervals between dotted lines are not constant. To this end, the data filtering unit (120) can perform a slice filter that slices point cloud data into groups at regular intervals in the direction of travel of the road, thereby preventing continuity detection between lanes from being interrupted. The regular interval at which the slice filter slices the point cloud data can be set as an initial value or can be set by the user as needed. For example, the regular interval can be 1 meter. Referring now to FIG. 5, point cloud data to which the slice filter has been applied will be described.
[0076]
[0077] FIG. 5 is a diagram illustrating an example of point cloud data to which a slice filter is applied by a data filtering unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0078] Referring to FIG. 5, it can be confirmed that the point cloud data is grouped by being cut at regular intervals by applying a slice filter by the data filtering unit (120) of the point cloud road lane extraction system (100) according to one embodiment of the present invention.
[0079]
[0080] Returning to Figure 1, the lane extraction unit (130) analyzes point cloud data to extract lanes. The lane extraction unit (130) may extract lanes by analyzing point cloud data stored in the point cloud data storage unit (110), or may extract lanes by analyzing point cloud data filtered through the data filtering unit (120).
[0081] In one embodiment, the lane extraction unit (130) can extract lanes by calculating the proximity state between points of point cloud data and using the K-means algorithm to group adjacent points, segmenting traffic surface marking areas on the road, calculating the minimum bounding box of the segmented traffic surface marking areas, and applying a linear extraction algorithm to separate linear components. For example, the lane extraction unit (130) can extract lanes by separating linear components using a linear extraction algorithm such as RANSAC (RANdom SAmple consensus). RANSAC is an algorithm that removes noise from a dataset and predicts a model, and is a method widely used in image processing. The K-means algorithm is an algorithm that performs clustering by setting the sum of the squares of the distances between the centroid of each group and the data objects within the group as a cost function and updating the group to which each data object belongs in the direction that minimizes this function value, and is an algorithm that can be used when grouping adjacent points. The algorithm used by the lane extraction unit (130) of the point cloud road lane extraction system (100) according to one embodiment of the present invention to extract lanes is not limited to the K-means algorithm, and any algorithm that can be used to group adjacent points may be used. Referring now to FIGS. 6A and 6B , point cloud data grouped by the lane extraction unit will be described.
[0082]
[0083] FIG. 6a and FIG. 6b are diagrams illustrating an example of point cloud data grouped by a lane extraction unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0084] Referring to Figures 6a and 6b, it can be confirmed that lanes within a road are grouped together, and road signs are grouped together. The lane extraction unit (130) extracts the grouped lanes.
[0085]
[0086] Returning to FIG. 1, the continuous detection unit (140) creates a continuous detection object in the first lane extracted by the lane extraction unit (130) and detects whether the continuous detection object overlaps with the second lane, thereby determining continuity between the first and second lanes. The process by which the continuous detection unit (140) determines continuity will be described in detail with reference to FIG. 7 below.
[0087]
[0088]
[0089] *Figure 7 is a drawing specifically illustrating a process of determining continuity by a continuous detection unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0090] Referring to FIG. 7, point cloud data filtered (S310) through the data filtering unit (120) of the point cloud road lane extraction system (100) according to one embodiment of the present invention is grouped (S320) in the lane extraction unit (130) and the first lane (11) is extracted (S330). Then, a continuous detection object (1) is generated in the first lane (11) in the continuous detection unit (140) (S340). The continuous detection unit (140) determines the continuity between the first lane (11) and the second lane (12) using the continuous detection object (1) (S150). In one embodiment, the generated continuous detection object (1) may have the shape of a triangle with one vertex located within the first lane (11) and a base opposite to the vertex facing in the direction of travel of the road. In the embodiment illustrated in FIG. 7, one of the vertices of the continuous detection object (1) is positioned at the center of the first lane (11), and the bottom opposite to the vertex is illustrated to face in the direction of travel of the road. At this time, the height of the continuous detection object (1) can be set to correspond to the spacing between lanes, and the length of the bottom of the continuous detection object (1) can be set to correspond to the range of the exit angle of the lane. In one embodiment, the setting values such as the height or the length of the bottom of the continuous detection object (1) can be set according to the specifications of the Traffic Surface Marking Installation and Management Manual (National Police Agency / www.police.go.kr). For example, according to the Traffic Surface Marking Installation and Management Manual, in the case of the specifications of the dotted line of the lane, the painted length is described as 300 cm to 1,000 cm for general roads and highways, the empty length is described as 1.0 to 2.0 times the painted length, and the width is described as 10 cm to 15 cm. Accordingly, the height of the continuous detection object (1) can be set to 300 cm to 2,000 cm, which is the same as the empty length of the lane, and the length of the bottom of the continuous detection object (1) can be set to 20 cm to 45 cm, which is 2 to 3 times the width of the lane.The continuous detection unit (140) detects whether the second lane (12) and the continuous detection object (1) overlap to determine the continuity between the first lane (11) and the second lane (12). For example, if the second lane (12) and the continuous detection object (1) overlap as in (a) or (b), the first lane (11) and the second lane (12) can be determined to be continuous, and if the second lane (12) does not exist at all as in (c) or the second lane (12) deviates from the continuous detection object (1) as in (d), so that the second lane (12) and the continuous detection object (1) do not overlap, the first lane (11) and the second lane (12) can be determined to be not continuous. Referring to FIGS. 8a and 8b below, the appearance of a continuous detection object created by a continuous detection unit (140) to detect continuity between lanes and the result of detecting continuity will be described.
[0091]
[0092] FIG. 8a and FIG. 8b are drawings illustrating an example of a continuous detection unit of a point cloud road lane extraction system according to one embodiment of the present invention generating a continuous detection object to detect continuity between lanes.
[0093] Referring to FIGS. 8A and 8B , the continuous detection unit (140) of the point cloud road lane extraction system (100) according to one embodiment of the present invention creates a continuous detection object for each lane to detect continuity between lanes. Each continuous detection object has a triangular shape, with one vertex positioned at the center of each lane, and the base opposite to the vertex faces the direction of travel of the road.
[0094]
[0095] FIG. 9a and FIG. 9b are diagrams illustrating an example of a result of detecting continuity between lanes by a continuous detection unit of a point cloud road lane extraction system according to one embodiment of the present invention.
[0096] Referring to FIGS. 9A and 9B , the continuous detection unit (140) of the point cloud road lane extraction system (100) according to one embodiment of the present invention shows the results of detecting continuity between lanes using continuous detection objects such as those shown in FIGS. 8A and 8B . In FIGS. 9A and 9B , lanes detected to be continuous with each other are grouped and displayed as numbers.
[0097]
[0098] Returning to FIG. 1, the missing interpolation unit (150) completes the connected line between the lanes of the road by granting continuity to the missing lanes based on the continuity between the lanes including the first and second lanes determined by the continuous detection unit (140). In an actual road, lanes may be erased or the bright light in a tunnel may prevent the lidar sensor from properly recognizing the lanes, resulting in the lanes not being detected and being omitted. That is, as shown in FIG. 9a, the result of detecting the continuity between lanes by the continuous detection unit (140) is that, although the lanes are actually connected, as in numbers 3 and 9, the tunnel section is not recognized when passing through the tunnel, and thus the lanes are determined to be separate, non-continuous lanes. In one embodiment, the missing interpolation unit (150) can complete a connected line between lanes of a road by inferring the line of mutually missing lanes by considering the height, bending angle, etc. of the lanes based on the continuity between the lanes determined by the continuous detection unit (140). Hereinafter, with reference to FIGS. 10a and 10b, the result of completing a connected line between lanes by the missing interpolation unit will be described.
[0099]
[0100] FIG. 10a and FIG. 10b are diagrams illustrating an example of a result of a missing interpolation unit of a point cloud road lane extraction system according to one embodiment of the present invention completing a connected line between lanes.
[0101] Referring to FIGS. 10A and 10B , the missing interpolation unit (150) of the point cloud road lane extraction system (100) according to one embodiment of the present invention provides continuity to missing lanes based on the continuity between lanes, thereby completing a connected line between road lanes. It can be seen in FIGS. 10A and 10B that even sections in tunnel sections where lanes were not detected are provided with continuity and connected by a light solid line.
[0102]
[0103] Fig. 11 is a flowchart for explaining a point cloud road lane extraction method according to one embodiment of the present invention.
[0104] As illustrated in FIG. 11, a point cloud road lane extraction method according to one embodiment of the present invention may include steps S410 to S450.
[0105] Step S410 is a step of storing point cloud (PCL) data collected on the road in the point cloud data storage unit (110).
[0106] Step S420 is a step of filtering the stored point cloud data in the data filtering unit (120). In one embodiment, the data filtering unit (120) may perform filtering by batch-shifting the signal intensity values of the stored point cloud data and amplifying and trimming the deviations between the signal intensity values. In another embodiment, the data filtering unit (120) may perform a drive near filter to extract only the point cloud data within a predetermined distance based on a driving line that is converted into a line by gathering points at the left and right points when the stored point cloud data is collected on the road. In addition, in another embodiment, the data filtering unit (120) may perform a slice filter to group the point cloud data on which the drive near filter has been performed by slicing the point cloud data into groups at predetermined intervals in the direction of travel of the road.
[0107] Step S430 is a step in which the lane extraction unit (130) analyzes point cloud data to extract lanes. In one embodiment, the lane extraction unit (130) can extract lanes by calculating the proximity between points of the point cloud data, using a K-means algorithm to group adjacent points, segmenting traffic surface marking areas on the road, calculating the minimum bounding box of the segmented traffic surface marking areas, and applying a linear extraction algorithm to segment linear components.
[0108] Step S440 is a step in which the continuous detection unit (140) creates a continuous detection object in the extracted first lane and detects whether the continuous detection object overlaps the second lane, thereby determining the continuity between the first lane and the second lane. In one embodiment, the continuous detection object may have a triangular shape with one vertex located within the first lane and a base opposite to the vertex facing in the direction of travel of the road. At this time, the height of the continuous detection object may be set to correspond to the spacing between lanes, and the length of the base of the continuous detection object may be set to correspond to the range of the lane's exit angle. In one embodiment, the setting values such as the height or the length of the base of the continuous detection object (1) may be set according to the specifications of the Traffic Surface Marking Installation and Management Manual (National Police Agency / www.police.go.kr). For example, according to the traffic surface marking installation and management manual, in the case of the specifications of the dotted line of the lane, the painted length is 300 cm to 1,000 cm for general roads and highways, the empty length is 1.0 to 2.0 times the painted length, and the width is 10 cm to 15 cm. Accordingly, the height of the continuous detection object (1) can be set to 300 cm to 2,000 cm, which is the same as the empty length of the lane, and the length of the bottom of the continuous detection object (1) can be set to 20 cm to 45 cm, which is 2 to 3 times the width of the lane. In addition, in one embodiment, the second lane is located in front of the first lane in the direction of travel of the road, and the continuous detection unit (140) can determine the continuity between the first lane and the second lane by detecting whether the bottom of the triangle of the continuous detection object overlaps the second lane.
[0109] Step S450 is a step of completing a linear connection between lanes of a road by providing continuity to the missing lanes based on the continuity between lanes including the first lane and the second lane in the missing interpolation unit (150).
[0110]
[0111] The point cloud road lane extraction method according to one embodiment of the present invention described above has been described with reference to the flowchart presented in the drawings. For simplicity, the method has been depicted and described as a series of blocks. However, the present invention is not limited to the order of the blocks. Some blocks may occur in a different order or simultaneously with other blocks than depicted and described herein, and various other branches, flow paths, and block orders that achieve the same or similar results may be implemented. Furthermore, not all of the depicted blocks may be required to implement the method described herein.
[0112]
[0113] Meanwhile, in the description referring to FIG. 11, each step may be further divided into additional steps or combined into fewer steps, depending on the implementation of the present invention. Furthermore, some steps may be omitted as needed, and the order of steps may be changed. Furthermore, even if other details are omitted, the contents of FIGS. 1 to 10b may be applied to the contents of FIG. 11. Furthermore, the contents of FIG. 11 may be applied to the contents of FIGS. 1 to 10b.
[0114]
[0115] FIG. 12 is a block diagram illustrating a computer system for implementing a point cloud road lane extraction method according to one embodiment of the present invention.
[0116] Referring to FIG. 12, a computer system (1000) may include at least one of a processor (1010), a memory (1030), an input interface device (1050), an output interface device (1060), and a storage device (1040) that communicate via a bus (1070). The computer system (1000) may further include a communication device (1020) coupled to a network. The processor (1010) may be a central processing unit (CPU), or a semiconductor device that executes instructions stored in the memory (1030) or the storage device (1040). The memory (1030) and the storage device (1040) may include various forms of volatile or non-volatile storage media. For example, the memory may include a read-only memory (ROM) and a random access memory (RAM). In embodiments of the present disclosure, the memory may be located internally or externally to the processor, and the memory may be connected to the processor via various known means. Memory is a variety of volatile or non-volatile storage media, and may include, for example, read-only memory (ROM) or random access memory (RAM).
[0117] Accordingly, embodiments of the present invention may be implemented as a computer-implemented method or as a non-transitory computer-readable medium storing computer-executable instructions. In one embodiment, when executed by a processor, the computer-readable instructions may perform a method according to at least one aspect of the present disclosure.
[0118] The communication device (1020) can transmit or receive wired or wireless signals.
[0119] In addition, the method according to the embodiment of the present invention may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium.
[0120] The computer-readable medium may include program commands, data files, data structures, etc., either singly or in combination. The program commands recorded on the computer-readable medium may be specially designed and configured for embodiments of the present invention, or may be known and usable by those skilled in the art of computer software. The computer-readable recording medium may include a hardware device configured to store and execute the program commands. For example, the computer-readable recording medium may be a magnetic medium such as a hard disk, a floppy disk, and a magnetic tape, an optical medium such as a CD-ROM or a DVD, a magneto-optical medium such as a floptical disk, a ROM, a RAM, a flash memory, etc. The program commands may include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer through an interpreter, etc.
[0121]
[0122] For reference, components according to embodiments of the present invention may be implemented in the form of software or hardware such as a digital signal processor (DSP), a field programmable gate array (FPGA), or an application specific integrated circuit (ASIC), and may perform certain roles.
[0123] However, 'components' are not limited to software or hardware, and each component may be configured to reside on an addressable storage medium or configured to trigger one or more processors.
[0124] Thus, as an example, components include components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables.
[0125] Components and the functionality provided within those components may be combined into a smaller number of components or further separated into additional components.
[0126]
[0127] Meanwhile, it will be understood that each block of the flowchart drawings and combinations of the flowchart drawings can be performed by computer program instructions. These computer program instructions can be installed on a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, so that the instructions executed by the processor of the computer or other programmable data processing equipment create a means for performing the functions described in the flowchart block(s). The computer program instructions can also be installed on a computer or other programmable data processing equipment, so that a series of operational steps are performed on the computer or other programmable data processing equipment to create a computer-executable process, so that the instructions executed by the computer or other programmable data processing equipment can also provide steps for performing the functions described in the flowchart block(s).
[0128] Additionally, each block may represent a module, segment, or portion of code that contains one or more executable instructions for performing a specific logical function(s). It should also be noted that in some alternative implementation examples, the functions described in the blocks may occur out of order. For example, two blocks depicted in succession may actually be executed substantially concurrently, or the blocks may sometimes be executed in reverse order, depending on their respective functions.
[0129]
[0130] The term 'part' or 'module' used in this embodiment means a software or hardware component such as an FPGA or ASIC, and the 'part' or 'module' performs certain roles. However, the 'part' or 'module' is not limited to software or hardware. The 'part' or 'module' may be configured to be on an addressable storage medium and may be configured to play one or more processors. Thus, as an example, the 'part' or 'module' includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functionality provided within the components and '~sub-parts' or '~modules' may be combined into a smaller number of components and '~sub-parts' or '~modules' or further separated into additional components and '~sub-parts' or '~modules'. In addition, the components and '~sub-parts' or '~modules' may be implemented to regenerate one or more CPUs within the device or secure multimedia card.
[0131]
[0132] Although the present invention has been described above with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various modifications and changes may be made to the present invention without departing from the spirit and scope of the present invention as set forth in the claims below.
Claims
1. A point cloud data storage unit that stores point cloud (PCL) data collected on the road; A lane extraction unit that extracts lanes by analyzing the above point cloud data; and A point cloud road lane extraction system characterized by including a continuous detection unit that creates a continuous detection object in the first lane extracted from the lane extraction unit and detects whether the continuous detection object overlaps with the second lane, thereby determining continuity between the first lane and the second lane.
2. In paragraph 1, The above continuous detection object is, A point cloud road lane extraction system characterized in that one vertex is located within the first lane and the base opposite to the vertex has the shape of a triangle facing in the direction of travel of the road.
3. In paragraph 2, The above second lane is, Located in front of the first lane in the direction of travel of the above road, The above continuous detection unit, A point cloud road lane extraction system characterized in that it determines continuity between the first lane and the second lane by detecting whether the base of the triangle of the continuous detection object overlaps the second lane.
4. In paragraph 1, Further comprising a data filtering unit for filtering point cloud data stored in the above point cloud data storage unit, The above lane extraction unit, A point cloud road lane extraction system characterized in that it extracts lanes by analyzing point cloud data filtered through the above data filtering unit.
5. In paragraph 4, The above data filtering unit, A point cloud road lane extraction system characterized in that filtering is performed by batch-shifting signal intensity values of point cloud data stored in the point cloud data storage unit and amplifying and trimming deviations between the signal intensity values.
6. In paragraph 4, The above data filtering unit, A point cloud road lane extraction system characterized in that it performs a drive near filter to extract only point cloud data within a predetermined distance based on a driving line converted into a line by collecting points at left and right points when point cloud data stored in the point cloud data storage unit is collected on a road.
7. In paragraph 6, The above data filtering unit, A point cloud road lane extraction system characterized in that it performs a slice filter that groups point cloud data that has undergone the above drive near filter by slicing the data into groups at regular intervals in the direction of travel of the road.
8. In paragraph 1, A point cloud road lane extraction system characterized in that it further includes a missing interpolation unit that completes a connected line between lanes of a road by providing continuity to missing lanes based on continuity between lanes including the first lane and the second lane determined by the above continuous detection unit.
9. In paragraph 1, The above lane extraction unit, A point cloud road lane extraction system characterized in that it extracts lanes by using a K-means algorithm to group adjacent points by calculating the proximity state between points of the above point cloud data, to divide traffic surface marking areas on a road, to calculate the minimum bounding box of the divided traffic surface marking areas, and to apply a linear extraction algorithm to divide linear components.
10. Point cloud data storage step for storing point cloud (PCL) data collected on the road; A lane extraction step for extracting lanes by analyzing the above point cloud data; and A point cloud road lane extraction method, characterized in that it includes a continuous detection step of creating a continuous detection object in the first lane extracted in the lane extraction step and detecting whether the continuous detection object overlaps with the second lane, thereby determining continuity between the first lane and the second lane.
Citation Information
Patent Citations
Display control device and display control program
JP2021024556A
Lane departure warning system and method for warning thereof
KR101508357B1
Wall finishing device for piping
KR1020220132315A
Method for constructing retaining wall on the sea using buoyancy and retaining wall assembly for the same
KR102332459B1
