KR20190014283A - Apparatus for manufacturing lane information and method thereof - Google Patents

Abstract

An apparatus and a method for generating lane information are disclosed. The apparatus for generating lane information comprises: an extracting unit for extracting three-dimensional color points from image data photographed from a moving object; a filtering unit for extracting a three-dimensional color point correspond to a lane from the three-dimensional color points by filtering the three-dimensional color points extracted by the extracting unit; a lane polygon generating unit for generating a lane polygon by grouping the three-dimensional color points filtered by the filtering unit; and a lane generating unit for generating lane information by using the lane polygon generated by the lane polygon generating unit.

Description

[0001] APPARATUS FOR MANUFACTURING LANE INFORMATION AND METHOD THEREOF [0002]

The present invention relates to an apparatus and method for generating lane information, and more particularly, to a lane information generating apparatus and method for generating lane information of a road using image data acquired from a moving object.

The vehicle terminal uses a GPS (Global Positioning System) to track the position of the vehicle in operation, displays it on a road map, and provides information related to the route to the destination.

The vehicle terminal generally provides a map using a two-dimensional image, and recently, a map is provided using a three-dimensional image. Because three-dimensional images provide information similar to running on real roads, realistic displays are needed rather than two-dimensional images. The three-dimensional image realistically displays the elevation, intersection, etc. of the road, and displays the number of lanes, pedestrian crossing, and lane information displayed on the road in detail.

Background Art [0002] The background art of the present invention is disclosed in Korean Patent Laid-Open Publication No. 10-2013-0039605 (Feb.

An object of the present invention is to provide a lane information generating apparatus and method for extracting a three-dimensional RGB point cloud based on an image acquired from a moving object and generating a lane using the extracted three-dimensional RGB point cloud.

An apparatus for generating lane information according to an aspect of the present invention includes: an extracting unit for extracting a three-dimensional color point from image data photographed from a moving object; A filtering unit for filtering a three-dimensional color point extracted by the extraction unit to extract a three-dimensional color point corresponding to a lane among the three-dimensional color points; A lane polygon generation unit for clustering the three-dimensional color points filtered by the filtering unit to generate a lane polygon; And a lane generator for generating lane information using the lane polygon generated by the lane polygon generator.

The filtering unit extracts a three-dimensional color point corresponding to the lane using the red value, the green value, and the blue value of the three-dimensional color point extracted by the extracting unit, and the filtering unit extracts Dimensional color point corresponding to the lane is extracted according to whether the RGB value is included within the set range by comparing the Red value, Green value and Blue value of the three-dimensional color point with the predetermined setting range, respectively .

The filtering unit of the present invention is characterized in that when the Red value, the Green value and the Blue value of the 3D color point extracted by the extracting unit are included in the setting range corresponding to the white lane, Dimensional color point extracted by the extracting unit is included in a setting range corresponding to the yellow lane, the red color value, the green color value, and the blue color value of the three- Dimensional color point extracted by the three-dimensional color point recognizing unit 3 as a three-dimensional color point corresponding to the yellow lane.

The lane polygon generation unit may include: a point cloud generation unit for generating a three-dimensional color point cloud in units of lanes by grouping the three-dimensional color points corresponding to the lane into lanes; A cube generation unit for generating cubes in each of the three-dimensional color points of the three-dimensional color point cloud generated by the point cloud generation unit and clustering the generated cubes; And a lane polygon detecting unit for detecting a lane polygon based on the coordinates of the cube clustered by the cube generating unit.

The cube of the present invention is characterized in that the width is generated so as to be included within a lane width range.

The point cloud generator of the present invention is characterized by grouping three-dimensional color points based on a correlation of whether cubes can overlap with each other.

The point cloud generating unit of the present invention is characterized in that the non-clustered three-dimensional color point among three-dimensional color points is recognized as noise and filtered.

The lane polygon detecting unit of the present invention is characterized in that a lane polygon is detected by using at least one of the upper left coordinates and the lower right coordinates of the bottom surface of the cube clustered by the cube generating unit.

The lane generator of the present invention is characterized by extracting a lane shape using the center point of the long side surface of the lane polygon generated by the lane polygon generation unit.

A method of generating lane information according to an aspect of the present invention includes extracting a 3D color point from image data photographed from a moving object; Extracting a three-dimensional color point corresponding to a lane among the three-dimensional color points by filtering the three-dimensional color point extracted by the filtering unit; The lane polygon generation unit clustering the three-dimensional color points filtered by the filtering unit to generate a lane polygon; And generating the lane information using the lane polygon generated by the lane polygon generation unit.

The filtering unit extracts a three-dimensional color point corresponding to a lane by using a red value, a green value, and a blue value of the three-dimensional color point extracted by the extracting unit, The red value, the green value and the blue value of the color point are respectively compared with the predetermined setting range, and the three-dimensional color point corresponding to the lane is extracted according to whether the RGB value is included within the set range.

The filtering unit of the present invention is characterized in that when the Red value, the Green value and the Blue value of the 3D color point extracted by the extracting unit are included in the setting range corresponding to the white lane, Dimensional color point extracted by the extracting unit is included in a setting range corresponding to the yellow lane, the red color value, the green color value, and the blue color value of the three- Dimensional color point extracted by the three-dimensional color point recognizing unit 3 as a three-dimensional color point corresponding to the yellow lane.

The step of generating the lane polygon according to the present invention includes the steps of: clustering three-dimensional color points corresponding to lanes into lanes to generate a three-dimensional color point cloud in lane-based units; Generating cubes in each of the three-dimensional color points of the three-dimensional color point cloud and grouping the generated cubes; And detecting a lane polygon based on coordinates of the clustered cube.

The cube of the present invention is characterized in that the width is generated so as to be included within a lane width range.

The clustering of the cubes of the present invention is characterized by clustering the cubes through correlation of whether the cubes can overlap with each other.

In the grouping of the cubes of the present invention, the three-dimensional color points that are not clustered through the correlation between the cubes are recognized as noise and filtered.

The step of detecting the lane polygon of the present invention is characterized in that the lane polygon is detected by using the left upper and left lower coordinates of the bottom surface of the clustered cube.

The lane generator of the present invention is characterized by extracting a lane shape using the center point of the long side surface of the lane polygon generated by the lane polygon generation unit.

The lane information generating apparatus and method according to an aspect of the present invention can generate a lane using an air vehicle, thereby greatly reducing the time and cost required for generating a lane and generating a map.

According to another aspect of the present invention, an apparatus and method for generating lane information can automatically improve lane information in an image acquired through a moving object, thereby greatly improving the efficiency and creativity of autonomous driving map production.

1 is a block diagram of a lane information generating apparatus according to an embodiment of the present invention.
2 is a block diagram of a lane polygon generation unit according to an embodiment of the present invention.
3 is a conceptual diagram illustrating a first-order filtering process using RGB information of a 3D RGB point cloud according to an exemplary embodiment of the present invention.
4 is a view showing a point cloud in a lane-based unit obtained as a result of a first-order filter according to an embodiment of the present invention.
5 is a view showing a point unit cube according to an embodiment of the present invention.
FIG. 6 is a view illustrating a clustered cube according to an embodiment of the present invention.
FIG. 7 is a diagram illustrating a method of forming a lane polygon using upper left coordinates and lower right coordinates of a clustered cube according to an exemplary embodiment of the present invention. Referring to FIG.
8 is a diagram illustrating a method of extracting a lane shape using a lane polygon center point according to an embodiment of the present invention.
9 is a flowchart illustrating a method of generating lane information according to an exemplary embodiment of the present invention.

Hereinafter, an apparatus and method for generating lane information according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings. In this process, the thicknesses of the lines and the sizes of the components shown in the drawings may be exaggerated for clarity and convenience of explanation. Further, the terms described below are defined in consideration of the functions of the present invention, which may vary depending on the user, the intention or custom of the operator. Therefore, definitions of these terms should be made based on the contents throughout this specification.

FIG. 1 is a block diagram of a lane information generating apparatus according to an embodiment of the present invention. FIG. 2 is a block diagram of a lane polygon generating unit according to an embodiment of the present invention. FIG. 4 is a view showing a point cloud of a lane-based unit obtained as a result of a first-order filter according to an embodiment of the present invention. FIG. FIG. 5 is a view illustrating a point unit cube according to an embodiment of the present invention. FIG. 6 is a view illustrating a clustered cube according to an embodiment of the present invention. FIG. 8 is a diagram illustrating a method of forming a lane polygon using the upper left coordinates and the lower right coordinates of the clustered cube according to an embodiment of the present invention. And a method of extracting a lane linearity.

Referring to FIG. 1, a lane information generating apparatus according to an embodiment of the present invention includes an extracting unit 10, a filtering unit 20, a lane polygon generating unit 30, and a lane generating unit 40.

The extraction unit 10 extracts a three-dimensional color point from the image data photographed from the moving object. Here, the mobile object may include a mobile MMS (Mobile Mapping System) vehicle or a flying object, and the flying object may include a dragon or the like.

That is, when the image data is inputted from various photographing equipments provided in the moving object, the extracting unit 10 processes the inputted image data to extract a three-dimensional color point.

The 3D color point is a point having a color in the image data, and has a red value, a green value, and a blue value.

The filtering unit 20 filters the three-dimensional color points extracted by the extraction unit 10 to extract three-dimensional color points corresponding to the lanes among the plurality of three-dimensional color points.

As described above, the three-dimensional color point has a red value, a green value, and a blue value, respectively. The filtering unit 20 calculates a red value, a green value, and a blue value of the three-dimensional color point extracted by the extracting unit 10, Value can be used to extract only the three-dimensional color point corresponding to the lane.

In this case, the filtering unit 20 compares the Red value, the Green value, and the Blue value of the three-dimensional color point extracted by the extracting unit 10 with the predetermined setting range, respectively, and determines that the Red value, the Green value, Dimensional color point corresponding to the lane is detected depending on whether or not it is included within the setting range.

As shown in FIG. 3, the setting range is a range of Red values, Green values, and Blue values that serve as criteria for determining whether a three-dimensional color point corresponds to a three-dimensional color point corresponding to a lane. A corresponding setting range, and a setting range corresponding to a yellow lane. This is because the lane is divided into a white lane and a yellow lane, and a lane can be identified as a white lane or a yellow lane based on a setting range corresponding to a white lane and a setting range corresponding to a yellow lane.

Accordingly, the filtering unit 20 determines whether or not three-dimensional color points corresponding to lanes 3, 4, 5, 6, and 7 corresponding to the lane of the three-dimensional color points are selected, based on whether the Red value, Green value, Extracted by the extraction unit 10 if the red value, the green value, and the blue value of the three-dimensional color point extracted by the extraction unit 10 are included in the setting range corresponding to the white lane, respectively, The three-dimensional color point is recognized as a three-dimensional color point corresponding to the white lane, and the Red value, the Green value, and the Blue value of the three-dimensional color point extracted by the extraction unit 10 are stored in the setting range corresponding to the yellow lane If it is included, the 3D color point extracted by the extraction unit 10 is recognized as a 3D color point corresponding to the yellow lane.

The lane polygon generation unit 30 clusters the three-dimensional color points filtered by the filtering unit 20 to generate a lane polygon.

2, the lane polygon generation unit 30 includes a point cloud generation unit 32, a cube generation unit 34, and a lane polygon detection unit 36. [

The point cloud generating unit 32 groups the three-dimensional color points corresponding to the lanes acquired by the filtering unit 20 in units of lanes to generate a three-dimensional color point cloud in a lane-based unit.

Referring to FIG. 4, the point cloud generating unit 32 groups the three-dimensional color points corresponding to the lanes acquired by the filtering unit 20 into lane-by-lane units to form a three-dimensional color point cloud do.

The cube generating unit 34 generates a cube in each of the three-dimensional color points of the three-dimensional color point cloud generated by the point cloud generating unit 32, and groups the generated cube.

The cube is formed in a hexahedron centered on a three-dimensional color point as shown in FIG. 5, and is formed such that its width is included within a lane width range. As the width of the cube is formed to be included within the lane width range, a lane polygon generated based on the coordinates of the cube can be generated corresponding to the width of the lane. This will be described later.

Since a plurality of three-dimensional color points exist in a three-dimensional color point cloud, cubes of three-dimensional color points can be clustered as shown in FIG.

In this case, the cube generating unit 34 groups the cubes of the three-dimensional color point according to the correlation between the cubes of the plurality of three-dimensional color points. That is, the cube generation unit 34 confirms whether the cubes of the adjacent three-dimensional color points overlap each other for the cubes of the three-dimensional color points, and groups the cubes of the three-dimensional color points overlapping each other. Accordingly, cubes of three-dimensional color points are clustered in lane-by-lane units.

Further, in the process of grouping the cubes of the three-dimensional color point according to the correlation between the cubes of the three-dimensional color point, as described above, the cube generation unit 34 generates a cube of an air point (air point) (Low points) are recognized as noise and filtered.

The lane polygon detecting unit 36 detects the lane polygon based on the coordinates of the cube clustered by the cube generating unit 34.

7, the lane polygon detecting unit 36 detects the lane polygon using the upper left coordinates and the lower right coordinates of the bottom surface of the cube clustered by the cube generating unit 34. [ In other words, the lane polygon detecting unit 36 extracts the upper left coordinates and the lower right coordinates of the cubes of these three-dimensional color points because each cube of the three-dimensional color point is formed within the range of the lane width as described above, So that a lane polygon corresponding to the lane can be formed by connecting the coordinates of the left upper end of the bottom surface of the floor and the right and left bottom coordinates.

As described above, the cubes are formed on the three-dimensional color points and the lane polygons are formed on the cubes. As a result, even if the aircraft is shaken due to the influence of the wind or the ground MMS vehicle is shaken due to vibration or running, You can filter the noise through the cube.

The lane generator 40 generates lane information using the lane polygon generated by the lane polygon generator 30. [

Referring to FIG. 8, the lane generator 40 detects center points of long sides for each polygon generated by the lane polygon generation unit 30, and extracts lane markers using these center points.

Hereinafter, a lane information generating method according to an embodiment of the present invention will be described with reference to FIG.

9 is a flowchart illustrating a method of generating lane information according to an exemplary embodiment of the present invention.

Referring to FIG. 9, the extracting unit 10 extracts a 3D color point having a Red value, a Green value, and a Blue value by image processing the input image data when the image data photographed from the moving object is input S10).

As the three-dimensional color point is extracted by the extraction unit 10, the filtering unit 20 converts the Red value, Green value, and Blue value of the three-dimensional color point extracted by the extraction unit 10 into a predetermined setting range Dimensional color point corresponding to the lane is detected according to whether the Red value, the Green value, and the Blue value are included within the predetermined setting range (S20).

At this time, if the Red value, Green value and Blue value of the three-dimensional color point extracted by the extracting unit 10 are included in the setting range corresponding to the white lane, the filtering unit 20 extracts And recognizes the three-dimensional color points as a three-dimensional color point corresponding to the white lane, and determines that the red value, the green value and the blue value of the three-dimensional color point extracted by the extraction section 10 are within the setting range corresponding to the yellow lane Dimensional color point extracted by the extraction unit 10 is recognized as a three-dimensional color point corresponding to the yellow lane.

As the three-dimensional color point corresponding to the lane is detected, the point cloud generating unit 32 groups the three-dimensional color points corresponding to the lane acquired by the filtering unit 20 into lane-by-lane units, A point cloud is generated (S30).

Then, the cube generating unit 34 generates a cubic cube of each of the three-dimensional color points of the three-dimensional color point cloud generated by the point cloud generating unit 32, and calculates the cube of the generated three- Clustering is performed according to the relationship (S40). In this case, the cube generating unit 34 confirms whether the cubes of the adjacent three-dimensional color points overlap each other with respect to each of the cubes of the three-dimensional color points, and groups the cubes of the three-dimensional color points overlapping each other. As a result, cubes of three-dimensional color points are clustered in lane-by-lane units. At this time, the cube generating unit 34 forms the cube of the three-dimensional color point so that its width is included within the lane width range.

In this process, the cube generating unit 34 recognizes air points or low points that are not clustered as noise and filters them.

The lane polygon detecting unit 36 extracts the upper left coordinates and the lower right coordinates of the bottom surface of the cube clustered by the cube creating unit 34 and extracts the upper left coordinates and the lower right coordinates of the bottom face of each of the cubes The upper left coordinates are connected to each other and the lower right coordinates are connected to each other to detect a lane polygon (S50).

Next, the lane generator 40 detects center points of the short sides for each polygon generated by the lane polygon generation unit 30, and extracts lane lines corresponding to the positions of the lane polygons around these center points (S60).

As described above, the lane information generating apparatus and method according to the embodiment of the present invention can greatly reduce the time and expense for lane creation and map production by generating lanes using air vehicles.

In addition, the lane information generating apparatus and method according to an embodiment of the present invention can automatically improve the efficiency and creativity of autonomous driving map generation by automatically generating lane information from an image acquired through a flight.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation, I will understand. Accordingly, the true scope of the present invention should be determined by the following claims.

10:
20:
30: a lane polygon generation unit
32: Point Cloud Generator
34: Cube generation unit
36: Lane polygon detector
40: lane generator

Claims (18)

An extracting unit for extracting a three-dimensional color point from the image data photographed from the moving object;
A filtering unit for filtering a three-dimensional color point extracted by the extraction unit to extract a three-dimensional color point corresponding to a lane among the three-dimensional color points;
A lane polygon generation unit for clustering the three-dimensional color points filtered by the filtering unit to generate a lane polygon; And
And a lane generator for generating lane information using the lane polygon generated by the lane polygon generator.
2. The image processing apparatus according to claim 1, wherein the filtering unit extracts a three-dimensional color point corresponding to the lane using the red value, the green value, and the blue value of the three-dimensional color point extracted by the extracting unit, The Green value and the Blue value of the three-dimensional color point extracted by the three-dimensional color point are compared with the predetermined setting range, respectively, and the three-dimensional color point corresponding to the lane is extracted according to whether or not the RGB value is included within the set range Wherein the lane information generating device comprises:
3. The image processing apparatus as claimed in claim 2, wherein the filtering unit includes a filtering unit for filtering the extracted three-dimensional color points, if the red value, the green value and the blue value of the three-dimensional color point extracted by the extracting unit are included in the setting range corresponding to the white lane, Dimensional color point is recognized as a three-dimensional color point corresponding to a white lane, and when the Red value, the Green value and the Blue value of the three-dimensional color point extracted by the extraction unit are included in the setting range corresponding to the yellow lane And recognizes the three-dimensional color point extracted by the extraction unit as a three-dimensional color point corresponding to a yellow lane.
The apparatus of claim 1, wherein the lane polygon generation unit
A point cloud generator for generating a three-dimensional color point cloud in a lane-by-lane by grouping the three-dimensional color points corresponding to the lane into lanes;
A cube generation unit for generating cubes in each of the three-dimensional color points of the three-dimensional color point cloud generated by the point cloud generation unit and clustering the generated cubes; And
And a lane polygon detecting unit for detecting a lane polygon based on the coordinates of the cube clustered by the cube generating unit.
5. The apparatus of claim 4, wherein the cube is generated such that the width is included within a lane width range.
5. The apparatus of claim 4, wherein the point cloud generator clusters three-dimensional color points based on a correlation of whether cubes can be overlapped with each other.
7. The apparatus of claim 6, wherein the point cloud generating unit recognizes non-clustered three-dimensional color points among three-dimensional color points as noise and filters the three-dimensional color points.
5. The apparatus of claim 4, wherein the lane polygon detecting unit detects a lane polygon using at least one of a left upper end coordinate and a lower right end coordinate of a bottom surface of a cube clustered by the cube generating unit.
The lane information generating apparatus according to claim 1, wherein the lane-generating section extracts a lane-shaped line using the center point of the long side surface of the lane polygon generated by the lane polygon generating section.
Extracting a three-dimensional color point from the image data photographed from the moving object;
Extracting a three-dimensional color point corresponding to a lane among the three-dimensional color points by filtering the three-dimensional color point extracted by the filtering unit;
The lane polygon generation unit clustering the three-dimensional color points filtered by the filtering unit to generate a lane polygon; And
And the lane generator generates the lane information using the lane polygon generated by the lane polygon generation unit.
11. The image processing apparatus according to claim 10, wherein the filtering unit extracts a three-dimensional color point corresponding to the lane using the red value, the green value, and the blue value of the three-dimensional color point extracted by the extracting unit, Dimensional color point corresponding to the lane is extracted according to whether the RGB value is within the set range or not by comparing the Red value, Green value and Blue value of the three-dimensional color point, Lane information generation method.
12. The image processing apparatus according to claim 11, wherein the filtering unit includes a filtering unit for filtering the extracted three-dimensional color points, if the Red value, the Green value, and the Blue value of the 3D color point extracted by the extracting unit are included in the setting range corresponding to the white lane, Dimensional color point is recognized as a three-dimensional color point corresponding to a white lane, and when the Red value, the Green value and the Blue value of the three-dimensional color point extracted by the extraction unit are included in the setting range corresponding to the yellow lane Wherein the three-dimensional color point extracted by the extracting unit is recognized as a three-dimensional color point corresponding to a yellow lane.
11. The method of claim 10, wherein generating the lane polygon comprises:
Generating a three-dimensional color point cloud in a lane-by-lane unit by grouping the three-dimensional color points corresponding to the lane into lanes;
Generating cubes in each of the three-dimensional color points of the three-dimensional color point cloud and grouping the generated cubes; And
And detecting a lane polygon based on the coordinates of the clustered cube.
14. The method of claim 13, wherein the cube is generated such that the width is included within a lane width range.
14. The method of claim 13, wherein clustering the cubes comprises clustering cubes based on a correlation of whether the cubes can overlap with each other.
16. The method according to claim 15, wherein in the step of clustering the cubes, the three-dimensional color points that are not clustered through the correlation between the cubes are recognized as noise and filtered.
14. The method of claim 13, wherein the detecting of the lane polygon comprises detecting a lane polygon using left upper and right lower coordinates of the bottom surface of the clustered cube.
14. The lane information generating method according to claim 13, wherein the lane-generating unit extracts a lane-shaped line using the center point of the long side surface of the lane polygon generated by the lane polygon generating unit.

Images