CN115248448A - Laser radar-based road edge detection method, device, equipment and storage medium - Google Patents
Laser radar-based road edge detection method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN115248448A CN115248448A CN202211154557.1A CN202211154557A CN115248448A CN 115248448 A CN115248448 A CN 115248448A CN 202211154557 A CN202211154557 A CN 202211154557A CN 115248448 A CN115248448 A CN 115248448A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- road edge
- road
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
- G01S17/931—Lidar systems specially adapted for specific applications for anti-collision purposes of land vehicles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/89—Lidar systems specially adapted for specific applications for mapping or imaging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
Abstract
The embodiment of the invention provides a method, a device, equipment and a storage medium for detecting a road edge based on a laser radar, and relates to the technical field of vehicle detection. The method comprises the following steps: acquiring point cloud data detected by a plurality of laser radars on a vehicle, and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area taking the vehicle as a center; segmenting the target point cloud data to obtain a segmentation result; determining left and right road edge curves and a road center reference line according to the segmentation result; determining a key point set according to the left road edge curve, the right road edge curve and the road center reference line; and performing curve fitting for multiple times according to the key point set to obtain a target road edge curve. In the embodiment of the invention, the road edge is quickly detected according to the left and right road edge curves and the road center reference line determined by the segmentation result to obtain the key point set, and then curve fitting is carried out according to the key point set, so that the identification accuracy is improved, and support is provided for subsequent accurate and quick sensing.
Description
Technical Field
The present invention relates to the field of vehicle detection technologies, and in particular, to a method and an apparatus for detecting a road edge based on a laser radar, an electronic device, and a storage medium.
Background
With the increase of the quantity of urban automobiles and the rapid development of Advanced assisted driving technologies (Advanced assisted driving systems), advanced assisted driving technologies (Advanced assisted driving systems) are more and more widely applied. By means of high-precision distance measurement and the characteristic of wide environment application, the application of the laser radar in environment perception is more and more extensive. The road edge detection can provide an accurate travelable boundary, assist positioning and subsequent decision making, and simultaneously play a great role in reducing laser perception false detection and environmental noise interference.
In the prior art, a multi-line mechanical rotation radar scanning beam is generally adopted to scan sudden change in the height or beam angle direction, characteristics are manually extracted for identification, or the detection is realized by manually designing the characteristic of the sudden change caused by the road edge in a multi-view (aerial view or front view). The method has the disadvantages of higher requirement on computing capacity, higher and higher requirement on computing as the point cloud points increase, and slow road edge identification.
Disclosure of Invention
In view of the above, embodiments of the present invention are proposed in order to provide a lidar based road edge detection method, a corresponding lidar based road edge detection apparatus, an electronic device, and a storage medium that overcome or at least partially solve the above-mentioned problems.
In order to solve the above problems, an embodiment of the present invention discloses a method for detecting a road edge based on a laser radar, including:
the method comprises the steps of obtaining point cloud data detected by a plurality of laser radars on a vehicle, and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area with the vehicle as the center;
segmenting the target point cloud data to obtain a segmentation result;
determining left and right road edge curves and a road center reference line according to the segmentation result;
determining a key point set according to the left road edge curve, the right road edge curve and the road center reference line;
and performing curve fitting on the key point set for multiple times to obtain a target road edge curve.
Optionally, the segmentation result includes an object class, and determining the left and right road edge curves and the road center reference line according to the segmentation result includes:
determining left and right road edge curves in the target area according to the object type;
and determining the center lines of the left and right road edge curves and using the center lines as road center reference lines.
Optionally, the determining a set of key points according to the left and right road edge curves and the road center reference line includes:
determining key points at preset distances forward and backward respectively at two ends of the left road edge curve and the right road edge curve along the direction of the road center reference line;
determining whether an obstacle exists in a preset area taking the key point as a center;
if no obstacle exists in a preset area corresponding to two key points at the same end of the left road edge curve and the right road edge curve, determining central points of the two key points, and prolonging the road central reference line according to the central points;
circularly executing the above operations until an obstacle is in a preset area with the key point as the center;
and determining each determined key point as a key point set.
Optionally, performing multiple curve fitting on the set of key points to obtain a target road edge curve includes:
dividing the key point set into a plurality of subsets according to the positions of the key points;
respectively performing curve fitting on the key points in each subset;
and determining a target road edge curve according to the curve corresponding to each subset.
Optionally, before the segmenting the target point cloud data to obtain a segmentation result, the method further includes:
preprocessing the target point cloud data to level the target point cloud data;
projecting the preprocessed target point cloud data into the divided 2D grids;
the step of segmenting the target point cloud data to obtain segmentation results comprises the following steps:
determining object categories in the target point cloud data;
and segmenting the target point cloud data in the 2D mesh according to the object type to obtain a segmentation result.
Optionally, the determining the object class in the target point cloud data includes:
extracting high-dimensional features of the target point cloud data;
and optimizing the high-dimensional features to obtain the object type in the target point cloud data.
Optionally, the obtaining point cloud data detected by a plurality of laser radars on the vehicle, and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area with the vehicle as a center includes:
converting the acquired point cloud data into a vehicle coordinate system to obtain initial point cloud data;
and screening the initial point cloud data in a target area taking a vehicle as a center to obtain target point cloud data.
Correspondingly, the embodiment of the invention discloses a road edge detection device based on a laser radar, which comprises:
the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring point cloud data detected by a plurality of laser radars on a vehicle, and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area taking the vehicle as a center;
the data segmentation module is used for segmenting the target point cloud data to obtain a segmentation result;
the result determining module is used for determining the left road edge curve, the right road edge curve and the road center reference line according to the segmentation result;
the set determining module is used for determining a key point set according to the left and right road edge curves and the road center reference line;
and the curve determining module is used for performing curve fitting on the key point set for multiple times to obtain a target road edge curve.
Optionally, the segmentation result includes an object class, and the result determination module includes:
the left and right road edge curve determining submodule is used for determining left and right road edge curves in the target area according to the object type;
and the road center reference line determining submodule is used for determining the center lines of the left and right road edge curves and using the center lines as the road center reference lines.
Optionally, the set determining module includes:
the key point determining submodule is used for determining key points at preset distances forward and backward from two ends of the left road edge curve and the right road edge curve respectively along the direction of the road center reference line;
the judgment submodule is used for determining whether an obstacle exists in a preset area with the key point as the center;
the central point determining submodule is used for determining the central points of the two key points if no obstacles exist in a preset area corresponding to the two key points at the same end of the left road edge curve and the right road edge curve, and prolonging the road central reference line according to the central points;
a circulation submodule for circularly performing the above operations until there is an obstacle in a preset area centered on the key point;
and the key point set determining submodule is used for determining each determined key point as a key point set.
Optionally, the curve determining module comprises:
the subset dividing submodule is used for dividing the key point set into a plurality of subsets according to the positions of the key points;
a curve fitting submodule for performing curve fitting on the key points in each of the subsets respectively;
and the target road edge curve submodule is used for determining a target road edge curve according to the curve corresponding to each subset.
Optionally, the apparatus further comprises:
the leveling module is used for preprocessing the target point cloud data so as to level the target point cloud data;
the projection module is used for projecting the preprocessed target point cloud data into the divided 2D grids;
the data segmentation module comprises:
a category determination submodule for determining a category of an object in the target point cloud data;
and the segmentation submodule is used for segmenting the target point cloud data in the 2D grid according to the object categories to obtain segmentation results.
Optionally, the category determination sub-module includes:
the high-dimensional feature extraction unit is used for extracting high-dimensional features of the target point cloud data;
and the optimizing unit is used for optimizing the high-dimensional features to obtain object categories in the target point cloud data.
Optionally, the data obtaining module includes:
the external reference calibration sub-module is used for converting the acquired point cloud data into a vehicle coordinate system to obtain initial point cloud data;
and the data screening submodule is used for screening the initial point cloud data in a target area taking a vehicle as a center to obtain target point cloud data.
The embodiment of the invention has the following advantages: the method comprises the steps that point cloud data detected by a plurality of laser radars on a vehicle are obtained, and the point cloud data of the laser radars are screened to obtain target point cloud data in a target area with the vehicle as the center; dividing the target point cloud data to obtain a division result; determining left and right road edge curves and a road center reference line according to the segmentation result; determining a key point set according to the left road edge curve, the right road edge curve and the road center reference line; and performing curve fitting for multiple times according to the key point set to obtain a target road edge curve. In the embodiment of the invention, the road edge is quickly detected according to the left and right road edge curves and the road center reference line determined by the segmentation result to obtain the key point set, and then curve fitting is carried out according to the key point set, so that the identification accuracy is improved, and support is provided for subsequent accurate and quick perception.
Drawings
Fig. 1 is a flowchart illustrating steps of a laser radar-based road edge detection method according to an embodiment of the present invention;
FIG. 2 is a target region detection map provided by an embodiment of the present invention;
FIG. 3 is a schematic diagram of a road edge detection provided by an embodiment of the invention;
fig. 4 is a block diagram of a structure of a lidar-based road edge detection apparatus according to an embodiment of the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
At present, a multi-line mechanical rotary radar scanning beam is generally adopted to suddenly change in height or angle direction of the beam, characteristics are manually extracted for identification, or the detection is realized by manually designing the sudden change characteristics caused by the road edge in a multi-view (aerial view or front view). However, as the number of point cloud points increases, the calculation process increases, the labor cost increases, and the recognition rate of the variable road edge conditions such as turning at the intersection is not high.
One of the core ideas of the embodiment of the invention is to provide a method for detecting the road edge of the laser radar, wherein the dimensions of point cloud data are improved by using a model, then the optimization is carried out, and after the point cloud data are segmented, a road edge curve is rapidly determined by using the combination of points and lines, so that the problems of low road edge condition identification rate and complex calculation process are solved.
Referring to fig. 1, a flowchart illustrating steps of a laser radar-based road edge detection method according to an embodiment of the present invention is shown, where the method specifically includes the following steps:
The method comprises the steps of obtaining laser radar detection point cloud data containing position information on a vehicle, displaying all the point cloud data in a laser radar coordinate system, and screening the point cloud data to obtain target point cloud data in a target area with the vehicle as the center. The point cloud is a data set, and each point in the data set represents a set of X, Y, Z geometric coordinates and an intensity value that records the intensity of the return signal as a function of the object surface reflectivity. When these points are combined together, a point cloud, i.e., a collection of data points in space representing a 3D shape or object, is formed.
In the embodiment of the present invention, the step of obtaining point cloud data detected by a plurality of laser radars on a vehicle and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area with the vehicle as a center may include the following sub-steps:
step S1011, converting the acquired point cloud data into a vehicle coordinate system to obtain initial point cloud data;
and a substep S1012, screening the initial point cloud data in a target area taking the vehicle as the center to obtain target point cloud data.
When the vehicle-mounted three-dimensional reconstruction system and the sensing system of the laser radar work, external parameters (three rotation parameters and three translation parameters) of the laser radar need to be calibrated, and the calibration of the external parameters of the laser radar refers to a relative transformation relation of a measurement coordinate system of the laser radar relative to measurement coordinate systems of other sensors, namely a rotation-translation transformation matrix. In the embodiment Of the invention, the acquired point cloud data can be converted into a vehicle coordinate system from a laser radar coordinate system, so that the point cloud data is subjected to external reference calibration to obtain initial point cloud data, then a target area is determined according to a Region Of Interest (ROI), the initial point cloud data which does not belong to the target area range is removed, and the initial point cloud data which belongs to the target area range is determined as the target point cloud data.
For example, referring to fig. 2, if the preset radius of the ROI is 20 meters, a circle is drawn with the vehicle as the center and the radius R1 is 20 meters, and the circle is the target region, the point cloud data outside the circle is removed, only the point cloud data inside the circle is retained, and other methods for obtaining ranges such as the diameter or the area of the ROI may also be preset, which is not limited herein in the embodiment of the present invention.
And 102, segmenting the target point cloud data to obtain a segmentation result.
And segmenting the determined target point cloud data to obtain a segmentation result. The segmentation is carried out according to the object type of the target point cloud data.
In the embodiment of the present invention, before the step of segmenting the target point cloud data to obtain the segmentation result, the method may further include the following sub-steps:
the substep S1021 is used for preprocessing the target point cloud data so as to level the target point cloud data;
step S1022, projecting the preprocessed target point cloud data into the divided 2D grids;
in an embodiment of the present invention, the step of segmenting the target point cloud data to obtain a segmentation result may include the following sub-steps:
a substep S1023 of determining object classes in the target point cloud data;
and a substep S1024 of segmenting the target point cloud data in the 2D grid according to the object categories to obtain segmentation results.
And preprocessing the target point cloud data by using the network model so as to level the target point cloud data. The network model preprocessing refers to leveling a coordinate system and an object, acquiring the posture of the three-dimensional space object, and turning over the posture to enable the posture to be coincident with a coordinate axis. One skilled in the art will appreciate that there are many ways to achieve this leveling, and the embodiments of the present invention are not limited thereto. And projecting the preprocessed target point cloud data into the divided 2D grids, wherein the 2D grids are grids with preset sizes, and the sizes of the grids can be set according to actual requirements. The embodiment of the invention projects the target point cloud data of the three-dimensional object into the two-dimensional 2D grid to play a role in reducing the dimension, and the object can be rapidly segmented in the 2D grid by utilizing the combination of points and lines to obtain a segmentation result.
In an embodiment of the present invention, the step of determining the object class in the target point cloud data may include the following sub-steps:
a substep S10231 of extracting high-dimensional features of the target point cloud data;
and a substep S10232 of optimizing the high-dimensional features to obtain object categories in the target point cloud data.
The point net model can be used for extracting high-dimensional characteristics of the target point cloud data, and then the point model can be used for optimizing the high-dimensional characteristics to obtain all object types in the target point cloud data, such as object types of flower beds, pedestrians, vehicles, houses and the like. The point net model is a network model for classifying and segmenting 3D objects by using input point cloud information, and the point model is three-dimensionally reconstructed in a point cloud form based on a deep learning method to enhance the precision of the point cloud. For example, the current target point cloud data is 3-dimensional, can be promoted to 10-dimensional through a point net model, and then the promoted high-dimensional features are promoted by using the point model, wherein the promotion process is to make the object types clearer and facilitate segmentation. The embodiment of the invention carries out dimensionality improvement and optimization on the acquired point cloud data, and aims to obtain target point cloud data with higher dimensionality and clearer object type; the dimension for extracting the high-dimensional features may be set according to actual requirements, and the embodiment of the present invention is not limited herein.
In the embodiment of the invention, the target point cloud data can be preprocessed by utilizing a network model so as to level the target point cloud data, and then the target point cloud data of the three-dimensional space object is projected into a two-dimensional 2D grid; then extracting high-dimensional characteristics of the target point cloud data by using a point net model, and optimizing the high-dimensional characteristics by using the point model, so that the target point cloud data with higher dimensionality and clearer object type can be obtained, and the object type in the target point cloud data can be determined; and finally, in the 2D grid, segmenting the target point cloud data according to the object type by utilizing the combination of points and lines to obtain the target point cloud data corresponding to the object. For example, the 2D mesh is segmented along the flower bed to obtain target point cloud data corresponding to the flower bed.
And 103, determining the left road edge curve, the right road edge curve and the road center reference line according to the segmentation result.
Determining left and right road edge curves and a road center reference line according to the segmentation result, wherein,
the left road edge curve and the right road edge curve comprise a left road edge curve and a right road edge curve, and the left road edge curve and the right road edge curve can be line segments respectively; the left road edge curve can be the boundary line between the road and the left edge of the road, and the right road edge curve can be the boundary line between the road and the right edge of the road; the boundary line between the road and the road edge can be obtained according to the segmentation result; for example, referring to fig. 2, if the two sides of the road are flower beds, the boundary line between the flower bed and the road in the target area is obtained, after the target point cloud data corresponding to the flower bed is determined, the boundary line between the target point cloud data corresponding to the road and the target point cloud data corresponding to the flower bed on the left side is a left road edge curve, and the boundary line between the target point cloud data corresponding to the road and the target point cloud data corresponding to the flower bed on the right side is a right road edge curve.
In an embodiment of the present invention, the segmentation result includes an object class, and the step of determining the left and right road-edge curves and the road-center reference line according to the segmentation result may include the sub-steps of:
substep S1031, determining left and right road edge curves in the target area according to the object type;
and a substep S1032 of determining the center lines of the left and right road edge curves and using the center lines as the road center reference lines.
And determining left and right road edge curves in the target area according to the object types of the segmentation result, wherein the segmentation result comprises the object types, and the center lines of the left and right road edge curves are determined and serve as road center reference lines. For example, referring to fig. 2, after the object type is determined, and the object is determined to be an object at the edge of the road, such as a flower bed, a step, and the like; if the two sides of the road are the flower beds, acquiring an intersection line of the flower beds in the target area and the road, determining a left road edge curve and a right road edge curve, acquiring center lines of the left road edge curve and the right road edge curve, and determining the center lines as road center reference lines. And 104, determining a key point set according to the left and right road edge curves and the road center reference line.
And determining a key point set according to the determined left and right road edge curves and the road center reference line, wherein the key points are not in the target area.
In an embodiment of the present invention, the step of determining the set of key points according to the left and right road edge curves and the road center reference line may include the sub-steps of:
in the substep S1041, determining key points at preset distances forward and backward respectively at two ends of the left and right road edge curves along the direction of the road center reference line;
a substep S1042 of determining whether there is an obstacle in a preset region centered on the key point;
in the substep S1043, if there is no obstacle in the preset region corresponding to two key points at the same end of the left and right road edge curves, determining the center points of the two key points, and extending the road center reference line according to the center points;
a substep S1044 of cyclically performing the above operations until there is an obstacle in a preset area centered on the key point;
in sub-step S1045, each determined keypoint is determined as a set of keypoints.
According to the determined road center reference line, key points are determined at preset distances forward and backward from two ends of the left and right road edge curves respectively along the front and back directions of the road center reference line, then whether an obstacle exists or not is determined in a preset area taking the preset distance of the key points as the center as the radius, if no obstacle exists in the preset area corresponding to the two key points at the same end of the left and right road edge curves, the center points of the two key points are determined, the road center reference line is extended according to the center points, and the operations are executed in a circulating mode until an obstacle exists in the preset area taking the key points as the center. The road center reference line is used for restricting the direction of acquiring the key points, and the preset distance for acquiring the key points can be set according to actual requirements.
For example, referring to fig. 3, after determining the road center reference line, the upper left key point is determined at a preset distance R2 along the direction above the road center reference line above the left road edge curve, and then the search is performed in a preset area with the upper left key point as the center and R2 as the radius. And similarly, determining an upper right key point at a preset distance R2 above the right road edge curve along the direction above the road center reference line, searching in a preset area with the upper right key point as the center and R2 as the radius, judging whether an obstacle exists in the preset area of the upper left key point and the preset area of the upper right key point, if no obstacle exists, acquiring a central point between the upper left key point and the upper right key point, extending the road center reference line by using the central point, acquiring the key points again with the upper left key point and the upper right key point as the starting points, constraining the direction by the central reference line and the preset distance R2, and circularly executing the operations until the obstacle exists in the preset area with the key points as the center. Similarly, the lower left key point of the left road edge curve and the lower right key point of the right road edge curve are processed by the same method, because the next key point is determined according to the previous key point and the determined direction of the key points is constrained by the road center reference line, the method is applicable to various road edge conditions such as straight roads, curved roads and the like, and the problem of low recognition rate of variable road edge conditions such as intersection turning and the like is solved.
Each key point has a position and a direction belonging to the key point, and each determined key point can be determined as a key point set according to the position and the direction of the key point; for example, some key points are above the left road edge curve, some key points are below the left road edge curve, some key points are above the right road edge curve, and some key points are below the right road edge curve, and all these key points may be determined as the key point set.
And 105, performing curve fitting on the key point set for multiple times to obtain a target road edge curve.
And performing curve fitting on the key point set for multiple times to obtain a target road edge curve, wherein the curve fitting determines a plurality of key points according to the positions and the directions of the key points, and the obtained key points are all adjacent points. For example, a plurality of key points above and below the left road edge curve are obtained from the key point set, the key points are connected with the left road edge curve one by one to perform curve fitting, and the curve fitting can be performed multiple times to avoid errors.
In this embodiment of the present invention, the step of performing curve fitting on the set of key points for multiple times to obtain the target road edge curve may include the following sub-steps:
substep S1051, dividing the key point set into a plurality of subsets according to the positions of the key points;
a substep S1052 of curve fitting the keypoints in each of the subsets, respectively;
and a substep S1053 of determining a target road edge curve according to the curve corresponding to each subset.
And dividing the key point set into a plurality of subsets according to the positions and the directions of the key points, respectively carrying out curve fitting on the key points in each subset, and determining a target road edge curve according to the curve corresponding to each subset. For example, there are 400 keypoints in the keypoint set, and according to the positions and directions of the keypoints, it can be determined that there are 100 keypoints above the left road edge curve, 100 keypoints below the left road edge curve, 100 keypoints above the right road edge curve, and 100 keypoints below the right road edge curve; 100 key points above the curve of the left road edge can be divided into 5 subsets according to the adjacent front and back sequence, each subset has 20 key points, and the 20 key points are subjected to curve fitting according to positions to obtain the curve of the subset; the other 3 key points with 100 key points can obtain curves corresponding to the subsets respectively, the curve fitting can be set to be curve fitting for multiple times according to actual requirements, and accuracy of the curve fitting is improved. Fitting curves of 5 subsets above the left road edge curve, determining a road edge curve above the left road edge curve, similarly obtaining road edge curves below the left road edge curve, above the right road edge curve and below the right road edge curve, then performing curve fitting on the road edge curve above the left road edge curve and the road edge curve below the left road edge curve to obtain a target road edge curve on the left side of the road, performing curve fitting on the road edge curve above the right road edge curve and the road edge curve below the right road edge curve to obtain a target road edge curve on the right side of the road, wherein the target road edge curve can be a line segment.
The embodiment of the invention has the following advantages: the method comprises the steps of obtaining point cloud data detected by a plurality of laser radars on a vehicle, and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area with the vehicle as the center; preprocessing the target point cloud data to level the target point cloud data; projecting the preprocessed target point cloud data into the divided 2D grids; extracting high-dimensional features of target point cloud data; optimizing the high-dimensional features to obtain object categories in the target point cloud data; segmenting the target point cloud data in the 2D grid according to the object type to obtain a segmentation result; determining left and right road edge curves and a road center reference line according to the segmentation result; determining a key point set according to the left and right road edge curves and the road center reference line; and performing curve fitting for multiple times according to the key point set to obtain a target road edge curve. In the embodiment of the invention, the high-dimensional characteristics of the target point cloud data can be extracted, the dimensionality of the target point cloud data is improved, more accurate object categories are obtained, the target point cloud data can be segmented in the 2D grid, the dimensionality of the target point cloud data is reduced, the calculated amount of the subsequent steps is reduced, the road edges can be quickly detected according to the left and right road edge curves and the road center reference line determined by the segmentation result, the key point set is obtained, curve fitting is carried out according to the key point set, the identification accuracy is improved, and support is provided for subsequent accurate and quick sensing.
It should be noted that for simplicity of description, the method embodiments are shown as a series of combinations of acts, but those skilled in the art will recognize that the embodiments are not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the embodiments. Further, those of skill in the art will appreciate that the embodiments described in the specification are presently preferred and that no particular act is required to implement the embodiments of the invention.
Referring to fig. 4, a block diagram of a structure of a lidar-based road edge detection apparatus according to an embodiment of the present invention is shown, which may specifically include the following modules:
the data acquisition module 201 is configured to acquire point cloud data detected by a plurality of laser radars on a vehicle, and screen the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area with the vehicle as a center;
a data segmentation module 202, configured to segment the target point cloud data to obtain a segmentation result;
a result determining module 203, configured to determine left and right road edge curves and a road center reference line according to the segmentation result;
a set determining module 204, configured to determine a set of key points according to the left and right road edge curves and the road center reference line;
and a curve determining module 205, configured to perform multiple curve fitting on the set of key points to obtain a target road edge curve.
Optionally, the segmentation result includes an object class, and the result determination module includes:
the left and right road edge curve determining submodule is used for determining left and right road edge curves in the target area according to the object type;
and the road center reference line determining submodule is used for determining the center lines of the left and right road edge curves and using the center lines as the road center reference lines.
Optionally, the set determining module includes:
the key point determining submodule is used for determining key points at preset distances forward and backward from two ends of the left road edge curve and the right road edge curve respectively along the direction of the road center reference line;
the judgment submodule is used for determining whether an obstacle exists in a preset area with the key point as the center;
the central point determining submodule is used for determining the central points of the two key points if no obstacles exist in a preset area corresponding to the two key points at the same end of the left road edge curve and the right road edge curve, and prolonging the road central reference line according to the central points;
a circulation submodule for circularly performing the above operations until there is an obstacle in a preset area centered on the key point;
and the key point set determining submodule is used for determining each determined key point as a key point set.
Optionally, the curve determining module comprises:
the subset dividing submodule is used for dividing the key point set into a plurality of subsets according to the positions of the key points;
a curve fitting submodule for performing curve fitting on the key points in each of the subsets respectively;
and the target road edge curve submodule is used for determining a target road edge curve according to the curve corresponding to each subset.
Optionally, the apparatus further comprises:
the leveling module is used for preprocessing the target point cloud data so as to level the target point cloud data;
the projection module is used for projecting the preprocessed target point cloud data into the divided 2D grids;
the data segmentation module comprises:
the category determination sub-module is used for determining the category of the object in the target point cloud data;
and the segmentation submodule is used for segmenting the target point cloud data in the 2D grid according to the object categories to obtain segmentation results.
Optionally, the category determination sub-module includes:
the high-dimensional feature extraction unit is used for extracting high-dimensional features of the target point cloud data;
and the optimizing unit is used for optimizing the high-dimensional features to obtain the object types in the target point cloud data.
Optionally, the data obtaining module includes:
the external reference calibration submodule is used for converting the acquired point cloud data into a vehicle coordinate system to obtain initial point cloud data;
and the data screening submodule is used for screening the initial point cloud data in a target area taking a vehicle as a center to obtain target point cloud data.
For the device embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, refer to the partial description of the method embodiment.
An embodiment of the present invention further provides an electronic device, including:
the method comprises a processor, a memory and a computer program which is stored on the memory and can run on the processor, wherein when the computer program is executed by the processor, each process of the embodiment of the laser radar-based road edge detection method is realized, the same technical effect can be achieved, and the method is not repeated herein for avoiding repetition.
The embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when being executed by a processor, the computer program implements each process of the above-mentioned laser radar-based road edge detection method embodiment, and can achieve the same technical effect, and is not described herein again to avoid repetition.
The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, apparatus, or computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications of these embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including the preferred embodiment and all changes and modifications that fall within the true scope of the embodiments of the present invention.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrases "comprising one of \ 8230; \8230;" does not exclude the presence of additional like elements in a process, method, article, or terminal device that comprises the element.
The method for detecting a road edge based on a laser radar and the device for detecting a road edge based on a laser radar provided by the invention are described in detail, and specific examples are applied to explain the principle and the implementation mode of the invention, and the description of the examples is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A road edge detection method based on laser radar is characterized by comprising the following steps:
the method comprises the steps of obtaining point cloud data detected by a plurality of laser radars on a vehicle, and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area with the vehicle as the center;
segmenting the target point cloud data to obtain a segmentation result;
determining left and right road edge curves and a road center reference line according to the segmentation result;
determining a key point set according to the left road edge curve, the right road edge curve and the road center reference line;
and performing curve fitting on the key point set for multiple times to obtain a target road edge curve.
2. The method of claim 1, wherein the segmentation result comprises object classes, and wherein determining the left and right road-edge curves and the road-center reference line according to the segmentation result comprises:
determining left and right road edge curves in the target area according to the object type;
and determining the center lines of the left and right road edge curves and using the center lines as road center reference lines.
3. The method of claim 1, wherein determining a set of keypoints from the left and right road-edge curves and the road-center reference line comprises:
determining key points at preset distances forward and backward respectively at two ends of the left road edge curve and the right road edge curve along the direction of the road center reference line;
determining whether an obstacle exists in a preset area with the key point as the center;
if no obstacle exists in a preset area corresponding to two key points at the same end of the left road edge curve and the right road edge curve, determining the central points of the two key points, and extending the road central reference line according to the central points;
circularly executing the above operations until an obstacle is in a preset area with the key point as the center;
and determining each determined key point as a key point set.
4. The method of claim 1, wherein said performing a plurality of curve fits on said set of keypoints to obtain a target road-edge curve, comprises:
dividing the key point set into a plurality of subsets according to the positions of the key points;
respectively performing curve fitting on the key points in each subset;
and determining a target road edge curve according to the curve corresponding to each subset.
5. The method of claim 1, wherein prior to said segmenting the target point cloud data into segmentation results, further comprising:
preprocessing the target point cloud data to level the target point cloud data;
projecting the preprocessed target point cloud data into the divided 2D grids;
the step of segmenting the target point cloud data to obtain segmentation results comprises the following steps:
determining object categories in the target point cloud data;
and segmenting the target point cloud data in the 2D mesh according to the object type to obtain a segmentation result.
6. The method of claim 5, wherein the determining the object class in the target point cloud data comprises:
extracting high-dimensional features of the target point cloud data;
and optimizing the high-dimensional features to obtain the object type in the target point cloud data.
7. The method of claim 1, wherein the step of obtaining point cloud data detected by a plurality of lidar on a vehicle and screening the point cloud data of the plurality of lidar to obtain target point cloud data in a target area centered on the vehicle comprises:
converting the acquired point cloud data into a vehicle coordinate system to obtain initial point cloud data;
and screening the initial point cloud data in a target area taking a vehicle as a center to obtain target point cloud data.
8. A lidar-based road edge detection apparatus, the apparatus comprising:
the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring point cloud data detected by a plurality of laser radars on a vehicle and screening the point cloud data of the plurality of laser radars to obtain target point cloud data in a target area taking the vehicle as a center;
the data segmentation module is used for segmenting the target point cloud data to obtain a segmentation result;
the result determining module is used for determining the left road edge curve, the right road edge curve and the road center reference line according to the segmentation result;
the set determining module is used for determining a key point set according to the left and right road edge curves and the road center reference line;
and the curve determining module is used for performing curve fitting on the key point set for multiple times to obtain a target road edge curve.
9. An electronic device, comprising: a processor, a memory and a computer program stored on the memory and executable on the processor, the computer program, when executed by the processor, implementing the steps of the lidar based road edge detection method according to any of claims 1-7.
10. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, which computer program, when being executed by a processor, carries out the steps of the lidar based road edge detection method according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211154557.1A CN115248448B (en) | 2022-09-22 | 2022-09-22 | Laser radar-based road edge detection method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211154557.1A CN115248448B (en) | 2022-09-22 | 2022-09-22 | Laser radar-based road edge detection method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115248448A true CN115248448A (en) | 2022-10-28 |
| CN115248448B CN115248448B (en) | 2022-12-16 |
Family
ID=83699557
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211154557.1A Active CN115248448B (en) | 2022-09-22 | 2022-09-22 | Laser radar-based road edge detection method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115248448B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150269445A1 (en) * | 2014-03-19 | 2015-09-24 | Denso Corporation | Travel division line recognition apparatus and travel division line recognition program |
| US20180260636A1 (en) * | 2017-03-13 | 2018-09-13 | Baidu Online Network Technology (Beijing) Co., Ltd. | Obstacle detecting method and apparatus, device and storage medium |
| CN110068836A (en) * | 2019-03-20 | 2019-07-30 | 同济大学 | A kind of laser radar curb sensory perceptual system of intelligent driving electric cleaning car |
| CN111985322A (en) * | 2020-07-14 | 2020-11-24 | 西安理工大学 | Road environment element sensing method based on laser radar |
| CN112964264A (en) * | 2021-02-07 | 2021-06-15 | 上海商汤临港智能科技有限公司 | Road edge detection method and device, high-precision map, vehicle and storage medium |
| US20210256274A1 (en) * | 2020-02-14 | 2021-08-19 | Denso Corporation | Apparatus for estimating road parameter |
| CN114740493A (en) * | 2022-04-12 | 2022-07-12 | 天津航天机电设备研究所 | Road edge detection method based on multi-line laser radar |
-
2022
- 2022-09-22 CN CN202211154557.1A patent/CN115248448B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150269445A1 (en) * | 2014-03-19 | 2015-09-24 | Denso Corporation | Travel division line recognition apparatus and travel division line recognition program |
| US20180260636A1 (en) * | 2017-03-13 | 2018-09-13 | Baidu Online Network Technology (Beijing) Co., Ltd. | Obstacle detecting method and apparatus, device and storage medium |
| CN110068836A (en) * | 2019-03-20 | 2019-07-30 | 同济大学 | A kind of laser radar curb sensory perceptual system of intelligent driving electric cleaning car |
| US20210256274A1 (en) * | 2020-02-14 | 2021-08-19 | Denso Corporation | Apparatus for estimating road parameter |
| CN111985322A (en) * | 2020-07-14 | 2020-11-24 | 西安理工大学 | Road environment element sensing method based on laser radar |
| CN112964264A (en) * | 2021-02-07 | 2021-06-15 | 上海商汤临港智能科技有限公司 | Road edge detection method and device, high-precision map, vehicle and storage medium |
| CN114740493A (en) * | 2022-04-12 | 2022-07-12 | 天津航天机电设备研究所 | Road edge detection method based on multi-line laser radar |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115248448B (en) | 2022-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110263652B (en) | Laser point cloud data identification method and device | |
| CN107341453B (en) | Lane line extraction method and device | |
| CN113412505B (en) | Processing unit and method for ordered representation and feature extraction of a point cloud obtained by a detection and ranging sensor | |
| CN110674705A (en) | Small-sized obstacle detection method and device based on multi-line laser radar | |
| CN110663060B (en) | Method, device, system and vehicle/robot for representing environmental elements | |
| CN113874927A (en) | Parking detection method, system, processing device and storage medium | |
| CN112711034B (en) | Object detection method, device and equipment | |
| CN113569958B (en) | Laser point cloud data clustering method, device, equipment and medium | |
| CN111476242A (en) | Laser point cloud semantic segmentation method and device | |
| US20200191971A1 (en) | Method and System for Vehicle Detection Using LIDAR | |
| CN115100616A (en) | Point cloud target detection method and device, electronic equipment and storage medium | |
| CN114241448A (en) | Method and device for obtaining heading angle of obstacle, electronic equipment and vehicle | |
| TW202017784A (en) | Car detection method based on LiDAR by proceeding the three-dimensional feature extraction and the two-dimensional feature extraction on the three-dimensional point cloud map and the two-dimensional map | |
| CN115248448B (en) | Laser radar-based road edge detection method, device, equipment and storage medium | |
| US20230258813A1 (en) | LiDAR Free Space Data Generator and LiDAR Signal Processing Method Using Multi-Modal Noise Filtering Scheme | |
| CN114387293A (en) | Road edge detection method and device, electronic equipment and vehicle | |
| CN112712062A (en) | Monocular three-dimensional object detection method and device based on decoupling truncated object | |
| CN111338336B (en) | Automatic driving method and device | |
| CN117274651B (en) | Object detection method and device based on point cloud and computer readable storage medium | |
| EP4310789A1 (en) | Prediction method for target object, computer device, and storage medium | |
| CN117075135B (en) | Vehicle feature detection method, system, storage medium and electronic equipment | |
| CN116482646A (en) | Detection frame identification method and device | |
| CN117789160A (en) | Multi-mode fusion target detection method and system based on cluster optimization | |
| CN116883973A (en) | Point cloud target detection method and device and electronic equipment | |
| CN114994706A (en) | Obstacle detection method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |