A method and system for detecting and correcting misidentification of road markings

Abstract

This invention provides a method and system for detecting and correcting misidentification of road markings. The method includes: acquiring zebra crossing data and road marking data collected through crowdsourcing; obtaining zebra crossing clusters and road marking clusters based on the bounding rectangle coordinates of all zebra crossings and road markings using DBSCAN clustering; when different categories of coordinate points exist in the data clusters, the corresponding areas are identified as misidentified areas; clustering analysis of the coordinate points in the misidentified areas is performed based on the k-means algorithm to determine the traffic sign type corresponding to the coordinate points, and the misidentified data is corrected. This scheme can achieve automatic detection and correction of misidentified road markings, effectively improving the efficiency of detecting and filtering misidentified data and reducing the probability of missed detections.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving, and in particular relates to a method and system for detecting and correcting misidentification of road markings. Background Technology

[0002] In the field of autonomous driving, low-cost crowdsourced data collection vehicles are widely deployed to collect high-frequency, high-freshness data. However, this data collection is characterized by low equipment cost, poor collection accuracy, and massive data volume. Crowdsourced data often contains misidentified data, with misidentification of zebra crossings and ground markings being a common occurrence. Preprocessing is typically required to filter out and correct these misidentified ground markings and zebra crossings before further processing and optimization can proceed. Relying solely on manual visual inspection to filter misidentified items is inefficient and has a high rate of missed detections. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method and system for detecting and correcting misidentification of ground markings, which is used to solve the problems of low efficiency and high false negative rate in crowdsourced ground marking misidentification detection.

[0004] In a first aspect of the present invention, a method for detecting and correcting misidentification of road markings is provided, comprising:

[0005] Obtain zebra crossing data and ground marking data collected through crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markings, obtain zebra crossing clusters and ground marking clusters through DBSCAN clustering.

[0006] When there are coordinate points of different categories in a data cluster, the corresponding area is identified as a misidentified area.

[0007] Cluster analysis of misidentified area coordinates is performed based on the k-means algorithm to determine the type of traffic sign corresponding to the coordinates and correct the misidentified data.

[0008] In a second aspect of the present invention, a ground marking misidentification detection and correction system is provided, comprising:

[0009] The data clustering module is used to acquire zebra crossing data and ground marker data collected through crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markers, it uses DBSCAN clustering to obtain zebra crossing clusters and ground marker clusters.

[0010] The misidentification filtering module is used to identify the corresponding area as a misidentified area when there are coordinate points of different categories in the data cluster.

[0011] The analysis and judgment module is used to perform cluster analysis on the coordinate points of misidentified areas based on the k-means algorithm, determine the traffic sign type corresponding to the coordinate points, and correct the misidentified data.

[0012] In a third aspect of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor executes the computer program to implement the steps of the method as described in the first aspect of the present invention.

[0013] In a fourth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method provided in the first aspect of the present invention.

[0014] In this embodiment of the invention, based on the coordinates of zebra crossings and ground markers collected through crowdsourcing, misidentified areas are detected by DBSCAN clustering, and then the k-means algorithm is used to detect and correct the misidentified data. This solves the problems of low efficiency and high false negative rate of manual visual inspection in screening misidentified items, and realizes automatic detection, identification and correction of erroneous data, which greatly improves the efficiency of screening misidentified data, reduces the false negative rate and ensures the accuracy and reliability of crowdsourced data. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a method for detecting and correcting misidentifications of road markings according to an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of the circumscribed rectangle of a zebra crossing and ground markings provided in one embodiment of the present invention;

[0018] Figure 3 This is a schematic diagram of a cluster of misidentified points provided in one embodiment of the present invention;

[0019] Figure 4 This is a schematic diagram of a ground marking misidentification detection and correction system provided in one embodiment of the present invention;

[0020] Figure 5 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0022] It should be understood that the terms "comprising" and other similar expressions in the specification, claims, and accompanying drawings of this invention are intended to cover a non-exclusive inclusion, such as a process, method, system, or apparatus that includes a series of steps or units and is not limited to the listed steps or units. Furthermore, "first" and "second" are used to distinguish different objects and are not intended to describe a specific order.

[0023] Please see Figure 1 The present invention provides a flowchart illustrating a method for detecting and correcting misidentifications of road markings, comprising:

[0024] S101. Obtain zebra crossing data and ground marking data collected by crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markings, obtain zebra crossing clusters and ground marking clusters through DBSCAN clustering.

[0025] When crowdsourced lanes collect road markings, due to the limited accuracy of the equipment, the recognition results of different vehicles collecting the same road markings may vary. Usually, recognition errors or mistakes need to be judged manually. This implementation can automatically detect and correct misidentified data.

[0026] The ground markings are generally road arrow markings or pedestrian crossing warning signs, such as straight ahead and left turn markings. They usually appear near zebra crossings and are prone to identification errors with zebra crossings.

[0027] Specifically, the zebra crossings and ground markings collected by crowdsourcing are represented by the four vertices of the minimum bounding rectangle, and the data categories corresponding to the four coordinate points are labeled; the zebra crossings and ground markings observed at the same location are stacked by the corresponding four coordinate points to obtain a coordinate point set, and the point set is clustered by DBSCAN.

[0028] For example, such as Figure 2 As shown, in the crowdsourced zebra crossing and ground marking data, each zebra crossing and ground marking is represented by the smallest bounded rectangle composed of four points A, B, C, and D, and each point is represented by its coordinates (x, y, z). The category is represented by category numbers J and K, respectively, with category J data marked with 0 and category K data marked with 1.

[0029] DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is a density-based clustering algorithm that can divide high-density regions into clusters.

[0030] DBSCAN clustering is performed on the coordinates of all zebra crossings and ground markings at the same location to obtain clusters of all zebra crossings and ground markings. The Eps neighborhood value can be set to 8, and the minimum number of data points in each cluster can be 4.

[0031] like Figure 3 As shown, 1 and 2 are both clustering results. In 2, there are coordinate points with different labels, which are misidentified point clusters.

[0032] S102. When there are coordinate points of different categories in the data cluster, the corresponding area is determined to be a misidentified area.

[0033] Each data cluster may contain coordinate points with different types of labels. For example, if a certain type of data contains coordinate points with both 1 and 0 labels, then the cluster corresponding to that type of data is a region with misidentification.

[0034] Specifically, if the coordinate points in a data cluster belong to the same category, then the data category in the data cluster has been correctly identified. If all the labels in a certain category are 1 or all are 0, then it is considered normal data.

[0035] S103. Based on the k-means algorithm, perform cluster analysis on the coordinate points of the misidentified area, determine the traffic sign type corresponding to the coordinate points, and correct the misidentified data.

[0036] The k-means algorithm, or k-means clustering algorithm, randomly selects K objects as initial cluster centers, then calculates the distance between each coordinate point and each cluster center, assigning each coordinate point to the nearest cluster center. The k-means algorithm can further classify coordinate points within misidentified areas to determine the type of traffic sign, i.e., whether it is a zebra crossing or a ground marking.

[0037] Specifically, calculate the maximum number N in the k-means cluster; set the initial weight for each data point, and perform N-2 clustering operations on the coordinate points of the misidentified area; perform data analysis on each clustering result and calculate the weight value of each coordinate point in the subclass data; and determine the traffic sign type corresponding to the coordinate points of the misidentified area based on the sum of the weight values ​​of the coordinate points corresponding to zebra crossings and ground markings.

[0038] If the weight value of the coordinate point corresponding to the zebra crossing is equal to the weight value of the coordinate point corresponding to the ground mark, then the coordinate point corresponding to the current area (i.e. the misidentified point cluster area) is marked as abnormal data.

[0039] For example, calculate the maximum number N of k-means clusters. If the maximum number of data points in this region is n, then N = n / 8, rounding up to the nearest integer (7.9 becomes 7).

[0040] Set an initial weight W for each data point, with an initial value of W = [0,0]. This means that the data point may be a class 0 data point and it may be a class 1 data point. The initial value is set to 0 for each.

[0041] The total number of data points in the region is E. The region data is subjected to N-2 K-means clustering operations based on the coordinate points. The number of clusters in the clustering parameters increases from 2 to N.

[0042] In each clustering result, the data of each subclass is analyzed. If the number of data points in this subclass is *e*, the number of points with label 0 is *a*, and the number of points with label 1 is *b*, then the *w* value of the first point in this subclass is updated. Update in the following way: Update the W value for points in all subclasses;

[0043] Once all N clustering results have been analyzed, the W values ​​of all points can be considered updated. Calculate the sum of the W values ​​of the four points A, B, C, and D for each traffic sign, resulting in WT. The data type corresponding to the larger of the two weights in WT is the type of the traffic sign. If the two weights in WT are equal, it is considered outlier data.

[0044] In this embodiment, two clustering algorithms are used in combination with a voting and scoring mechanism to quickly detect misidentified zebra crossings or ground markings. This effectively solves the problems of low efficiency and high false negative rate of manual screening, greatly improves the screening efficiency and reduces the false negative rate, and can directly correct misidentified items.

[0045] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0046] Figure 4 This is a schematic diagram of a ground marking misidentification detection and correction system provided in an embodiment of the present invention. The system includes:

[0047] The data clustering module 410 is used to acquire zebra crossing data and ground marker data collected by crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markers, it uses DBSCAN clustering to obtain zebra crossing clusters and ground marker clusters.

[0048] Specifically, the zebra crossings and ground markings collected by crowdsourcing are represented by the four vertices of the minimum bounding rectangle, and the data categories corresponding to the four coordinate points are labeled.

[0049] Zebra crossings and ground markings observed at the same location are stacked using their corresponding four coordinate points to obtain a set of coordinate points, which are then subjected to DBSCAN clustering.

[0050] The misidentification filtering module 420 is used to determine the corresponding area as a misidentified area when there are coordinate points of different categories in the data cluster;

[0051] If the coordinate points in a data cluster are of the same category, then the data category in the data cluster is correctly identified.

[0052] The analysis and judgment module 430 is used to perform cluster analysis on the coordinate points of the misidentified area based on the k-means algorithm, determine the traffic sign type corresponding to the coordinate points, and correct the misidentified data.

[0053] Specifically, the analysis and determination module 430 includes:

[0054] The computational unit is used to calculate the maximum number N in a k-means cluster;

[0055] Clustering units are used to set the initial weight of each data point and perform N-2 clustering operations on the coordinates of misidentified areas.

[0056] The weight calculation unit is used to perform data analysis on each clustering result and calculate the weight value of each coordinate point in the subclass data.

[0057] The judgment unit is used to determine the type of traffic sign corresponding to the coordinate points of the misidentified area based on the weight values ​​of the coordinate points corresponding to the zebra crossing and the ground marking.

[0058] Furthermore, if the weight value of the coordinate point corresponding to the zebra crossing is equal to the weight value of the coordinate point corresponding to the ground mark, then the corresponding coordinate point in the current area is marked as abnormal data.

[0059] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the systems and modules described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0060] Figure 5 This is a schematic diagram of an electronic device according to an embodiment of the present invention. The electronic device is used for the detection and correction of misidentification of zebra crossings and ground arrows in crowdsourced data collection. Figure 5 As shown, the electronic device 5 of this embodiment includes: a memory 510, a processor 520, and a system bus 530. The memory 510 includes an executable program 5101 stored thereon. As those skilled in the art will understand, Figure 5The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0061] The following is combined Figure 5 A detailed introduction to each component of the electronic device:

[0062] The memory 510 can be used to store software programs and modules. The processor 520 executes various functional applications and data processing of the electronic device by running the software programs and modules stored in the memory 510. The memory 510 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device (such as cached data), etc. In addition, the memory 510 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0063] The memory 510 contains an executable program 5101 for a sign extraction method. This executable program 5101 can be divided into one or more modules / units, which are stored in the memory 510 and executed by the processor 520 to achieve functions such as zebra crossing and road marking misidentification detection. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, describing the execution process of the computer program 5101 in the electronic device 5. For example, the computer program 5101 can be divided into a data clustering module, a misidentification filtering module, and an analysis and judgment module.

[0064] The processor 520 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 510, and by calling data stored in the memory 510, it performs various functions and processes data, thereby monitoring the overall status of the electronic device. Optionally, the processor 520 may include one or more processing units; preferably, the processor 520 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, application programs, etc., and the modem processor mainly handles wireless communication. It is understood that the modem processor may also not be integrated into the processor 520.

[0065] System bus 530 is used to connect various functional components inside the computer, transmitting data, address, and control information. Its type can be, for example, PC I bus, ISA bus, or VESA bus. Instructions from processor 520 are transmitted to memory 510 via the bus, and memory 510 sends data back to processor 520. System bus 530 is responsible for data and instruction exchange between processor 520 and memory 510. Of course, system bus 530 can also connect to other devices, such as network interfaces and display devices.

[0066] In this embodiment of the invention, the executable program executed by the processing 520 of the electronic device includes:

[0067] Obtain zebra crossing data and ground marking data collected through crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markings, obtain zebra crossing clusters and ground marking clusters through DBSCAN clustering.

[0068] When there are coordinate points of different categories in a data cluster, the corresponding area is identified as a misidentified area.

[0069] Cluster analysis of misidentified area coordinates is performed based on the k-means algorithm to determine the type of traffic sign corresponding to the coordinates and correct the misidentified data.

[0070] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0071] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0072] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting and correcting misidentifications of road markings, characterized in that, include: Obtain zebra crossing data and ground marking data collected through crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markings, obtain zebra crossing clusters and ground marking clusters through DBSCAN clustering. In this method, the zebra crossings and ground markings collected by crowdsourcing are represented by the four vertices of the minimum bounding rectangle, and the data categories corresponding to the four coordinate points are labeled. Zebra crossings and ground markings observed at the same location are stacked using their corresponding four coordinate points to obtain a set of coordinate points, and then DBSCAN clustering is performed on the set of points. When there are coordinate points of different categories in a data cluster, the corresponding area is identified as a misidentified area. Cluster analysis of the coordinate points in the misidentified area is performed based on the k-means algorithm to determine the type of traffic sign corresponding to the coordinate points and correct the misidentified data.

2. The method according to claim 1, characterized in that, The step of determining the corresponding region as a misidentified region when there are coordinate points of different categories in the data cluster includes: If the coordinate points in a data cluster are of the same category, then the data category in the data cluster has been correctly identified.

3. The method according to claim 1, characterized in that, The specific steps for clustering the coordinates of misidentified areas using the k-means algorithm to determine the traffic sign type corresponding to the coordinates are as follows: Calculate the maximum number N in a k-means cluster; Set the initial weight for each data point, and perform N-2 clustering operations on the coordinates of the misidentified areas; For each clustering result, perform data analysis and calculate the weight value of each coordinate point in the sub-class data; Based on the weight values ​​of the coordinate points corresponding to zebra crossings and ground markings, determine the type of traffic sign corresponding to the coordinate points in the misidentified area.

4. The method according to claim 3, characterized in that, The method of determining the traffic sign type corresponding to the coordinate points of the misidentified area based on the sum of the weight values ​​of the coordinate points corresponding to the zebra crossing and the ground marking includes: If the weight value of the coordinate point corresponding to the zebra crossing is equal to the weight value of the coordinate point corresponding to the ground mark, then the corresponding coordinate point in the current area is marked as abnormal data.

5. A system for detecting and correcting misidentifications of road markings, characterized in that, include: The data clustering module is used to acquire zebra crossing data and ground marker data collected through crowdsourcing. Based on the bounding rectangle coordinates of all zebra crossings and ground markers, it uses DBSCAN clustering to obtain zebra crossing clusters and ground marker clusters. In this method, the zebra crossings and ground markings collected by crowdsourcing are represented by the four vertices of the minimum bounding rectangle, and the data categories corresponding to the four coordinate points are labeled. Zebra crossings and ground markings observed at the same location are stacked using their corresponding four coordinate points to obtain a set of coordinate points, and then DBSCAN clustering is performed on the set of points. The misidentification filtering module is used to determine the corresponding area as a misidentified area when there are different categories of coordinate points in the data cluster; the different categories of coordinate points refer to the coordinate points of zebra crossings or ground markings. The analysis and judgment module is used to perform cluster analysis on the coordinate points of misidentified areas based on the k-means algorithm, determine the traffic sign type corresponding to the coordinate points, and correct the misidentified data.

6. The system according to claim 5, characterized in that, The analysis and judgment module includes: The computational unit is used to calculate the maximum number N in a k-means cluster. Clustering units are used to set the initial weight of each data point and perform N-2 clustering operations on the coordinates of misidentified areas. The weight calculation unit is used to perform data analysis on each clustering result and calculate the weight value of each coordinate point in the subclass data. The judgment unit is used to determine the type of traffic sign corresponding to the coordinate points of the misidentified area based on the weight values ​​of the coordinate points corresponding to the zebra crossing and the ground marking.

7. An electronic 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 computer program, it implements the steps of the ground marking misidentification detection and correction method as described in any one of claims 1 to 4.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed, it implements the steps of the ground marking misidentification detection and correction method as described in any one of claims 1 to 4.

Images