Method and device for extracting guardrails from road point cloud
Through the combination of random sampling consistency algorithm and clustering algorithm, the problem of difficulty in extracting guardrail point cloud in the existing technology is solved, accurate extraction without prior information is achieved, and the complexity of extracting guardrail point cloud is reduced. .
Patent Information
- Application Number
- CN202010615878.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-06-30
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2040-06-30
AI Technical Summary
The prior art is difficult to accurately extract guardrail point clouds in large-scale road scenarios, and a priori position information or semantic rules are often required as basic conditions, which increases the difficulty of guardrail point cloud extraction.
The random sampling consistency algorithm and clustering algorithm are used to extract the guardrail point cloud from the original mobile laser scanning point cloud through preprocessing, translation, projection and screening steps, so as to achieve accurate extraction without location information.
It reduces the difficulty of extracting guardrail point clouds, and realizes accurate extraction of guardrail point clouds without prior information in large-scale road scenarios.
Smart Images
Figure CN111783648B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of three-dimensional data processing, and in particular to a method for extracting guardrails in a road point cloud, a computer-readable storage medium, a computer device, and a device for extracting guardrails in a road point cloud. Background Art
[0002] In the field of high-precision maps and autonomous driving, object detection and extraction in road environments are important and necessary.
[0003] In the related art, the methods for detecting and extracting guardrails in road point clouds mostly adopt image-based methods, point cloud-based methods, or methods based on both; however, these methods are difficult to accurately extract guardrail point clouds from large-scale road scene point cloud data; and, these methods often require prior location information or semantic rules as basic conditions before extraction, which increases the difficulty of extracting guardrail point clouds. Summary of the invention
[0004] The present invention aims to solve at least one of the technical problems in the above-mentioned technologies to a certain extent. To this end, one purpose of the present invention is to propose a method for extracting guardrails from road point clouds, which can accurately extract guardrail point clouds from large-scale road scene point clouds without prior information such as position information, thereby reducing the difficulty of guardrail point cloud extraction.
[0005] A second object of the present invention is to provide a computer-readable storage medium.
[0006] A third object of the present invention is to provide a computer device.
[0007] The fourth objective of the present invention is to provide a device for extracting guardrails in a road point cloud.
[0008] To achieve the above-mentioned purpose, the first aspect of the embodiment of the present invention proposes a method for extracting guardrails in road point clouds, comprising the following steps: obtaining an original mobile laser scanning point cloud, and preprocessing the original mobile laser scanning point cloud to generate an original point cloud; extracting road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud and a road plane model corresponding to the road plane point cloud, and vertically translating the road plane model, and filtering the first point cloud according to the translated road plane model to generate a second point cloud; projecting the second point cloud to a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and clustering the two-dimensional projection according to a clustering algorithm to generate multiple classes corresponding to the two-dimensional projection; screening the multiple classes to obtain a class corresponding to the guardrail, and extracting a three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail to complete the extraction of the guardrail in the road point cloud.
[0009] According to the method for extracting guardrails in road point clouds of an embodiment of the present invention, first, an original mobile laser scanning point cloud is obtained, and the original mobile laser scanning point cloud is preprocessed to generate an original point cloud; then, road plane points are extracted from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud and a road plane model corresponding to the road plane point cloud, and the road plane model is vertically translated, and the first point cloud is filtered according to the translated road plane model to generate a second point cloud; then, the second point cloud is projected onto a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and the two-dimensional projection is clustered according to a clustering algorithm to generate multiple classes corresponding to the two-dimensional projection; then, the multiple classes are screened to obtain a class corresponding to the guardrail, and a three-dimensional point cloud corresponding to the guardrail in the second point cloud is extracted according to the class corresponding to the guardrail to complete the extraction of the guardrail in the road point cloud; thereby, accurate extraction of guardrail point clouds from large-scale road scene point clouds is achieved without prior information such as position information, thereby reducing the difficulty of guardrail point cloud extraction.
[0010] In addition, the method for extracting guardrails from a road point cloud according to the above embodiment of the present invention may also have the following additional technical features:
[0011] Optionally, the original moving laser scanning point cloud is preprocessed, including: downsampling the original moving laser scanning point cloud; determining whether the height of any point in the downsampled original moving laser scanning point cloud is greater than a preset height threshold, and filtering out the point when the determination result is yes; and cutting the filtered original moving laser scanning point cloud according to a preset size to generate point cloud blocks of equal size.
[0012] Optionally, the clustering algorithm includes a DBSCAN density clustering algorithm.
[0013] Optionally, the multiple classes are screened to obtain the class corresponding to the guardrail, including: determining whether the points of any one of the multiple classes are less than a preset point threshold, and if the determination result is yes, excluding the class to obtain the class corresponding to the guardrail.
[0014] Optionally, the multiple classes are screened to obtain the class corresponding to the guardrail, including: performing line fitting on any one of the multiple classes to generate a straight line that conforms to the least squares solution based on all points of the class, and calculating the fitting error of the straight line that conforms to the least squares solution corresponding to each point, and calculating the average fitting error corresponding to the class based on the fitting error of each point; judging whether the average fitting error is greater than a preset fitting error threshold, and excluding the class when the judgment result is yes, to obtain the class corresponding to the guardrail.
[0015] Optionally, the multiple classes are screened to obtain the class corresponding to the guardrail, including: obtaining the reflection intensity of all points in any one of the multiple classes, and calculating the average reflection intensity of the points in the class based on the reflection intensities of all points; determining whether the average reflection intensity is not within a preset reflection intensity range, and if the determination result is yes, excluding the class to obtain the class corresponding to the guardrail.
[0016] To achieve the above-mentioned purpose, the second aspect of the present invention proposes a computer-readable storage medium, on which a program for extracting guardrails in road point clouds is stored. When the program for extracting guardrails in road point clouds is executed by a processor, the method for extracting guardrails in road point clouds as described above is implemented.
[0017] According to the computer-readable storage medium of an embodiment of the present invention, by storing a program for extracting guardrails in road point clouds, the processor implements the method for extracting guardrails in road point clouds as described above when executing the program for extracting guardrails in road point clouds, thereby accurately extracting guardrail point clouds from large-scale road scene point clouds without prior information such as position information, thereby reducing the difficulty of extracting guardrail point clouds.
[0018] In order to achieve the above-mentioned purpose, the third aspect of the present invention proposes a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that when the processor executes the program, a method for extracting guardrails in a road point cloud as described above is implemented.
[0019] According to the computer device of an embodiment of the present invention, the guardrail extraction program in the road point cloud is stored in the memory, so that when the processor executes the guardrail extraction program in the road point cloud, it implements the guardrail extraction method in the road point cloud as described above, thereby achieving accurate extraction of guardrail point clouds from large-scale road scene point clouds without prior information such as position information, thereby reducing the difficulty of guardrail point cloud extraction.
[0020] In order to achieve the above-mentioned purpose, the fourth aspect of the present invention proposes a device for extracting guardrails in road point clouds, including: a preprocessing module, the preprocessing module is used to obtain an original mobile laser scanning point cloud, and preprocess the original mobile laser scanning point cloud to generate an original point cloud; a filtering module, the filtering module is used to extract road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud and a road plane model corresponding to the road plane point cloud, and vertically translate the road plane model, and filter the first point cloud according to the translated road plane model to generate a second point cloud; a clustering module, the clustering module is used to project the second point cloud to a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and cluster the two-dimensional projection according to a clustering algorithm to generate multiple classes corresponding to the two-dimensional projection; an extraction module, the extraction module is used to screen the multiple classes to obtain a class corresponding to the guardrail, and extract a three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail to complete the extraction of the guardrail in the road point cloud.
[0021] According to the device for extracting guardrails from road point clouds of an embodiment of the present invention, a preprocessing module is set to obtain an original mobile laser scanning point cloud, and the original mobile laser scanning point cloud is preprocessed to generate an original point cloud; a filtering module is used to extract road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud and a road plane model corresponding to the road plane point cloud, and the road plane model is vertically translated, and the first point cloud is filtered according to the translated road plane model to generate a second point cloud; a clustering module is used to project the second point cloud to a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and cluster the two-dimensional projection according to a clustering algorithm to generate multiple classes corresponding to the two-dimensional projection; an extraction module is used to screen multiple classes to obtain a class corresponding to the guardrail, and extract a three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail, so as to complete the extraction of the guardrail in the road point cloud; thereby, the guardrail point cloud can be accurately extracted from a large-scale road scene point cloud without prior information such as position information, thereby reducing the difficulty of guardrail point cloud extraction.
[0022] In addition, the device for extracting guardrails from road point clouds according to the above embodiment of the present invention may also have the following additional technical features:
[0023] Optionally, the original moving laser scanning point cloud is preprocessed, including: downsampling the original moving laser scanning point cloud; determining whether the height of any point in the downsampled original moving laser scanning point cloud is greater than a preset height threshold, and filtering out the point when the determination result is yes; and cutting the filtered original moving laser scanning point cloud according to a preset size to generate point cloud blocks of equal size. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 A schematic diagram of a flow chart of a method for extracting guardrails from a road point cloud according to an embodiment of the present invention;
[0025] Figure 2 A schematic diagram of the effect of the guardrail extraction process according to an embodiment of the present invention;
[0026] Figure 3 4 is a block diagram of a device for extracting guardrails from a road point cloud according to an embodiment of the present invention. DETAILED DESCRIPTION
[0027] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.
[0028] In the related art, when extracting guardrail point clouds, it is necessary to use location information or rules as basic conditions, and it is difficult to accurately extract guardrail point clouds from large-scale road scene point cloud data; according to the method for extracting guardrails in road point clouds of an embodiment of the present invention, first, an original mobile laser scanning point cloud is obtained, and the original mobile laser scanning point cloud is preprocessed to generate an original point cloud; then, road plane points are extracted from the original point cloud according to a random sampling consistency algorithm to obtain the first point cloud, the road plane point cloud and the road plane model corresponding to the road plane point cloud, and the road plane model is vertically plane-planarized. The method comprises the following steps: shifting the guardrail and filtering the first point cloud according to the shifted road plane model to generate a second point cloud; then, projecting the second point cloud to a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and clustering the two-dimensional projection according to a clustering algorithm to generate a plurality of classes corresponding to the two-dimensional projection; then, screening the plurality of classes to obtain a class corresponding to the guardrail, and extracting a three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail, so as to complete the extraction of the guardrail in the road point cloud; thereby accurately extracting the guardrail point cloud from the large-scale road scene point cloud without prior information such as position information, and reducing the difficulty of guardrail point cloud extraction.
[0029] In order to better understand the above technical solution, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.
[0030] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0031] Figure 1 FIG. 1 is a flow chart of a method for extracting guardrails from a road point cloud according to an embodiment of the present invention. Figure 1 As shown, the method for extracting guardrails in the road point cloud includes the following steps:
[0032] S101, obtaining an original moving laser scanning point cloud, and preprocessing the original moving laser scanning point cloud to generate an original point cloud.
[0033] There are many ways to pre-process the original mobile laser scanning point cloud.
[0034] As an example, the original moving laser scanning point cloud is preprocessed, including: downsampling the original moving laser scanning point cloud; determining whether the height of any point in the downsampled original moving laser scanning point cloud is greater than a preset height threshold, and filtering out the point when the determination result is yes; cutting the filtered original moving laser scanning point cloud according to a preset size to generate point cloud blocks of equal size.
[0035] That is, first, the original mobile laser scanning point cloud (such as Figure 2 (a) is downsampled to reduce the number of points while maintaining the shape characteristics of the point cloud; then, it is determined whether the height of any point in the downsampled original mobile laser scanning point cloud is greater than a preset height threshold, and when the judgment result is yes, the point is filtered out. It can be understood that the road point cloud obtained by mobile laser scanning contains many high-rise buildings and tall tree crowns on the roadside; these points can be effectively excluded by the preset height Yunzhi; then, the filtered original mobile laser scanning point cloud is cut with a preset size to generate multiple point cloud blocks of equal size (such as Figure 2 (b)); furthermore, the preprocessing process can make the processed point cloud convenient for subsequent processing and reduce the computational complexity.
[0036] S102, extracting road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud, and a road plane model corresponding to the road plane point cloud, vertically translating the road plane model, and filtering the first point cloud according to the translated road plane model to generate a second point cloud.
[0037] That is to say, the road plane points are extracted from the original point cloud (which can be each point cloud block after preprocessing) according to the RANSAC (RANdom SAmple Consensus) random sampling consensus algorithm. It can be understood that the random sampling consensus algorithm can estimate the parameters of the mathematical model in an iterative manner from a set of observation data sets containing outliers, so as to eliminate the interference of erroneous samples to a certain extent; thus, through the random sampling consensus algorithm, the first point cloud (i.e., the remaining point cloud after filtering out the road plane points in the original point cloud), the road plane point cloud and the plane model corresponding to the road plane point cloud can be obtained; then, the road plane model is vertically translated, and after the translation, the interfering point cloud in the first point cloud whose height is outside the guardrail height range can be filtered out with the road plane model as the boundary; and then the second point cloud (such as Figure 2 (c)).
[0038] S103 , projecting the second point cloud onto a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and clustering the two-dimensional projection according to a clustering algorithm to generate a plurality of clusters corresponding to the two-dimensional projection.
[0039] That is, the second point cloud is projected onto a two-dimensional plane from a bird's-eye view. It can be understood that when the second point cloud is observed from a bird's-eye view, the road guardrail can be clearly identified. Therefore, the second point cloud is projected from a bird's-eye view to generate a two-dimensional projection, and the two-dimensional projection is clustered according to the clustering algorithm to generate multiple clusters corresponding to the two-dimensional projection (such as Figure 2 (d)).
[0040] There are many options for clustering algorithms.
[0041] As an example, the clustering algorithm includes DBSCAN (Density-Based Spatial Clustering of Applications with Noise) density clustering algorithm. It can be understood that DBSCAN is a density-based spatial clustering algorithm that is insensitive to the order of samples, does not need to know the number of clusters formed in advance, can find clusters of any shape, and can identify noise points at the same time.
[0042] S104, screening multiple classes to obtain the class corresponding to the guardrail, and extracting the three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail, so as to complete the extraction of the guardrail in the road point cloud.
[0043] That is to say, after clustering, the clustering results (i.e., multiple classes) are screened to exclude non-guardrail classes, so that the class corresponding to the guardrail can be finally obtained, and the three-dimensional point cloud corresponding to the guardrail in the second point cloud is proposed based on the class corresponding to the guardrail, thereby completing the extraction of the guardrail in the road point cloud.
[0044] There are many ways to filter multiple classes to obtain the class corresponding to the guardrail.
[0045] As an example, multiple classes are filtered to obtain the class corresponding to the guardrail, including:
[0046] It is determined whether the number of points of any one of the multiple classes is less than a preset point threshold, and if the determination result is yes, the class is excluded to obtain the class corresponding to the guardrail.
[0047] It can be understood that in the original mobile laser scanning point cloud collected by the lidar, the farther the point cloud is from the road, the sparser the distribution of its points; thus, the class farther away from the road contains fewer points; therefore, by comparing the number of points corresponding to each class with the point count threshold, the class far away from the road can be effectively excluded.
[0048] As another example, multiple classes are filtered to obtain the class corresponding to the guardrail, including:
[0049] Performing line fitting on any one of the multiple classes, so as to generate a straight line that meets the least squares solution according to all points of the class, and calculating the fitting error of the straight line that meets the least squares solution corresponding to each point, and calculating the average fitting error corresponding to the class according to the fitting error of each point;
[0050] It is determined whether the average fitting error is greater than a preset fitting error threshold, and if the determination result is yes, the class is excluded to obtain the class corresponding to the guardrail.
[0051] It can be understood that under the bird's-eye view projection, the guardrail is in the shape of a regular, thin straight line. Therefore, a line fitting operation is performed on each class, and a straight line that conforms to the least squares solution is fitted based on the coordinates of all points in each class. Then, the fitting error of the fitting result (that is, the straight line that conforms to the least squares solution) to each point is calculated; and the directly corresponding average fitting error is calculated based on the fitting error of each point; then, it is determined whether the average fitting error is greater than the preset fitting error threshold. If so, the class is excluded.
[0052] As another example, multiple classes are filtered to obtain the class corresponding to the guardrail, including:
[0053] Obtain the reflection intensity of all points in any class among multiple classes, and calculate the average reflection intensity of the points in the class based on the reflection intensity of all points;
[0054] It is determined whether the average reflection intensity is not within a preset reflection intensity range, and if the determination result is yes, the class is excluded to obtain the class corresponding to the guardrail.
[0055] It can be understood that due to the special role played by road guardrails in the road environment, their materials are different from those of general objects. As a result, the reflection intensity of road guardrails captured by the lidar is different from the reflection intensity generated by general objects. Therefore, the average reflection intensity corresponding to all points in each class is calculated, and it is determined whether the average reflection intensity is within the preset reflection intensity range. If not, the class is excluded, so as to screen the class corresponding to the guardrail well.
[0056] It should be noted that after screening multiple classes through the above steps, there may be a small number of classes that do not belong to the guardrail. Most of these misjudged classes are small in area but large in point density. Due to their small area and high point density, they are misjudged. Therefore, we can use the smallest box aligned with the coordinate axis to include all points in a class, and add restrictions on the box size to exclude this type of class.
[0057] To better illustrate, Figure 2 (e) shows the class of the guardrail after screening. Figure 2 (f) is the three-dimensional point cloud corresponding to the guardrail extracted according to the class corresponding to the guardrail.
[0058] In summary, according to the method for extracting guardrails in road point clouds according to an embodiment of the present invention, first, an original mobile laser scanning point cloud is obtained, and the original mobile laser scanning point cloud is preprocessed to generate an original point cloud; then, road plane points are extracted from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud, and a road plane model corresponding to the road plane point cloud, and the road plane model is vertically translated, and the first point cloud is filtered according to the translated road plane model to generate a second point cloud; then, the second point cloud is projected onto a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and the two-dimensional projection is clustered according to a clustering algorithm to generate multiple classes corresponding to the two-dimensional projection; then, the multiple classes are screened to obtain a class corresponding to the guardrail, and a three-dimensional point cloud corresponding to the guardrail in the second point cloud is extracted according to the class corresponding to the guardrail to complete the extraction of the guardrail in the road point cloud; thereby, accurate extraction of guardrail point clouds from large-scale road scene point clouds is achieved without prior information such as location information, thereby reducing the difficulty of guardrail point cloud extraction.
[0059] In order to implement the above embodiment, an embodiment of the present invention proposes a computer-readable storage medium, on which a program for extracting guardrails in a road point cloud is stored. When the program for extracting guardrails in a road point cloud is executed by a processor, the method for extracting guardrails in a road point cloud as described above is implemented.
[0060] According to the computer-readable storage medium of an embodiment of the present invention, by storing a program for extracting guardrails in road point clouds, the processor implements the method for extracting guardrails in road point clouds as described above when executing the program for extracting guardrails in road point clouds, thereby accurately extracting guardrail point clouds from large-scale road scene point clouds without prior information such as position information, thereby reducing the difficulty of extracting guardrail point clouds.
[0061] In order to implement the above-mentioned embodiment, an embodiment of the present invention proposes a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that when the processor executes the program, a method for extracting guardrails in a road point cloud as described above is implemented.
[0062] According to the computer device of an embodiment of the present invention, the guardrail extraction program in the road point cloud is stored in the memory, so that when the processor executes the guardrail extraction program in the road point cloud, it implements the guardrail extraction method in the road point cloud as described above, thereby achieving accurate extraction of guardrail point clouds from large-scale road scene point clouds without prior information such as position information, thereby reducing the difficulty of guardrail point cloud extraction.
[0063] In order to implement the above embodiment, the embodiment of the present invention proposes a device for extracting guardrails in a road point cloud, such as Figure 3 As shown, the device for extracting guardrails in the road point cloud includes: a preprocessing module 10, a filtering module 20, a clustering module 30 and an extraction module 40.
[0064] The preprocessing module 10 is used to obtain the original mobile laser scanning point cloud and preprocess the original mobile laser scanning point cloud to generate an original point cloud;
[0065] The filtering module 20 is used to extract road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud, and a road plane model corresponding to the road plane point cloud, and to vertically translate the road plane model, and to filter the first point cloud according to the translated road plane model to generate a second point cloud;
[0066] The clustering module 30 is used to project the second point cloud onto a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and cluster the two-dimensional projection according to a clustering algorithm to generate a plurality of clusters corresponding to the two-dimensional projection;
[0067] The extraction module 40 is used to screen multiple classes to obtain the class corresponding to the guardrail, and extract the three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail to complete the extraction of the guardrail in the road point cloud.
[0068] In some embodiments, the original moving laser scanning point cloud is preprocessed, including: downsampling the original moving laser scanning point cloud; determining whether the height of any point in the downsampled original moving laser scanning point cloud is greater than a preset height threshold, and filtering out the point when the determination result is yes; cutting the filtered original moving laser scanning point cloud according to a preset size to generate point cloud blocks of equal size.
[0069] It should be noted that the above Figure 1 The description of the method for extracting guardrails in the road point cloud is also applicable to the device for extracting guardrails in the road point cloud, which will not be repeated here.
[0070] In summary, according to the device for extracting guardrails in road point clouds according to an embodiment of the present invention, a preprocessing module is set to obtain the original mobile laser scanning point cloud, and the original mobile laser scanning point cloud is preprocessed to generate the original point cloud; the filtering module is used to extract road plane points from the original point cloud according to the random sampling consistency algorithm to obtain the first point cloud, the road plane point cloud and the road plane model corresponding to the road plane point cloud, and the road plane model is vertically translated, and the first point cloud is filtered according to the translated road plane model to generate a second point cloud; the clustering module is used to project the second point cloud to a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and the two-dimensional projection is clustered according to the clustering algorithm to generate multiple classes corresponding to the two-dimensional projection; the extraction module is used to screen multiple classes to obtain the class corresponding to the guardrail, and extract the three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail, so as to complete the extraction of the guardrail in the road point cloud; thereby, the guardrail point cloud can be accurately extracted from the large-scale road scene point cloud without prior information such as position information, thereby reducing the difficulty of guardrail point cloud extraction.
[0071] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0072] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0073] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0075] It should be noted that in the claims, any reference signs placed between brackets shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claim. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention may be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.
[0076] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0077] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
[0078] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0079] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0080] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.
[0081] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.
[0082] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.
Claims
1. A method for extracting guardrails from a road point cloud, characterized in that: The following steps are involved: Acquire an original moving laser scanning point cloud, and pre-process the original moving laser scanning point cloud to generate an original point cloud; Extracting road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud, and a road plane model corresponding to the road plane point cloud, vertically translating the road plane model, and filtering the first point cloud according to the translated road plane model to generate a second point cloud; Projecting the second point cloud onto a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and clustering the two-dimensional projection according to a clustering algorithm to generate a plurality of clusters corresponding to the two-dimensional projection; Screening the multiple classes to obtain a class corresponding to the guardrail, and extracting a three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail, so as to complete the extraction of the guardrail in the road point cloud; The multiple classes are filtered to obtain the class corresponding to the guardrail, including: Obtaining the reflection intensity of all points in any one of the multiple classes, and calculating the average reflection intensity of the points in the class based on the reflection intensity of all points; Determine whether the average reflection intensity is not within a preset reflection intensity range, and if the determination result is yes, exclude the class to obtain a class corresponding to the guardrail; After screening the multiple classes to obtain the class corresponding to the guardrail, the method further includes: A minimum box aligned with the coordinate axis is used to include all points in a class, and restrictions are added to the size of the minimum box to remove classes that are mistakenly identified as guardrails.
2. The method for extracting guardrails in a road point cloud as claimed in claim 1, characterized in that: The original mobile laser scanning point cloud is preprocessed, including: Downsampling the original mobile laser scanning point cloud; Determine whether the height of any point in the downsampled original mobile laser scanning point cloud is greater than a preset height threshold, and if the judgment result is yes, filter out the point; The filtered raw mobile laser scanning point cloud is cut according to the preset size to generate point cloud blocks of equal size.
3. The method for extracting guardrails in a road point cloud as claimed in claim 1, characterized in that: The clustering algorithm includes a DBSCAN density clustering algorithm.
4. A computer-readable storage medium, characterized in that: A guardrail extraction program in a road point cloud is stored thereon, and when the guardrail extraction program in a road point cloud is executed by a processor, a guardrail extraction method in a road point cloud as described in any one of claims 1 to 3 is implemented.
5. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for extracting guardrails in a road point cloud as described in any one of claims 1 to 3 is implemented.
6. A device for extracting guardrails from a road point cloud, characterized in that: include: A preprocessing module, the preprocessing module is used to obtain an original mobile laser scanning point cloud and preprocess the original mobile laser scanning point cloud to generate an original point cloud; A filtering module, the filtering module is used to extract road plane points from the original point cloud according to a random sampling consistency algorithm to obtain a first point cloud, a road plane point cloud, and a road plane model corresponding to the road plane point cloud, and to vertically translate the road plane model, and to filter the first point cloud according to the translated road plane model to generate a second point cloud; A clustering module, the clustering module is used to project the second point cloud onto a two-dimensional plane from a bird's-eye view to generate a two-dimensional projection, and cluster the two-dimensional projection according to a clustering algorithm to generate a plurality of classes corresponding to the two-dimensional projection; An extraction module, the extraction module is used to screen the multiple classes to obtain a class corresponding to the guardrail, and extract a three-dimensional point cloud corresponding to the guardrail in the second point cloud according to the class corresponding to the guardrail, so as to complete the extraction of the guardrail in the road point cloud; The multiple classes are filtered to obtain the class corresponding to the guardrail, including: Obtaining the reflection intensity of all points in any one of the multiple classes, and calculating the average reflection intensity of the points in the class based on the reflection intensity of all points; Determine whether the average reflection intensity is not within a preset reflection intensity range, and if the determination result is yes, exclude the class to obtain a class corresponding to the guardrail; After screening the multiple classes to obtain the class corresponding to the guardrail, the method further includes: A minimum box aligned with the coordinate axis is used to include all points in a class, and restrictions are added to the size of the minimum box to remove classes that are mistakenly identified as guardrails.
7. The device for extracting guardrails from a road point cloud as claimed in claim 6, characterized in that: The original mobile laser scanning point cloud is preprocessed, including: Downsampling the original mobile laser scanning point cloud; Determine whether the height of any point in the downsampled original mobile laser scanning point cloud is greater than a preset height threshold, and if the judgment result is yes, filter out the point; The filtered raw mobile laser scanning point cloud is cut according to the preset size to generate point cloud blocks of equal size.