A method for determining a missing vector element and an electronic device
By acquiring target nodes and displaying vector features in high-precision maps, and matching reference vector features with standard images, the problem of missing vector features in high-precision maps is solved, thereby achieving accuracy of navigation paths and effectiveness of autonomous driving.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 安徽蔚来智驾科技有限公司
- Filing Date
- 2023-06-12
- Publication Date
- 2026-06-26
AI Technical Summary
The existence of missing vector features in existing high-precision maps affects navigation accuracy, especially when important routes such as ramps are missing, leading to problems such as autonomous driving missing service areas.
By acquiring the target nodes and their displayed vector features in the target image, obtaining reference vector features using the standard image, and matching them, the reference vector features with target type attributes are filtered out, and the distance is calculated or the missing vector features are determined using a hidden Markov model.
Automatically and effectively identify missing vector features, improve navigation accuracy, avoid using display vector features with low matching degree as substitutes, and ensure the correctness of navigation paths.
Smart Images
Figure CN116659531B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of navigation technology, specifically providing a method for determining missing vector elements and an electronic device. Background Technology
[0002] Autonomous driving relies heavily on high-precision maps. However, due to the high difficulty and cost of creating high-precision maps, map providers often struggle to provide perfectly complete maps, resulting in missing map elements in existing high-precision maps. For example, in the acquisition and production of high-precision maps, ramps are prioritized less than highways and main roads, leading to a large number of missing ramps in some high-precision maps. However, in some cases, these missing ramps are crucial for route planning and matching SD (Standard Definition) to HD (High-Precision) routes. For instance, when an SD route passes through a service area, the HD map, lacking information on the service area and its on / off ramps, will use a relatively high-matching route from the HD map as the target output, failing to automatically and effectively identify the missing route. This can result in matching an HD route on the highway, where the autonomous driving system might miss the service area due to reliance on the high-precision map, leading to a poor user experience. Summary of the Invention
[0003] The present invention aims to solve the above-mentioned technical problem, namely, to solve the problem that missing vector elements exist in existing high-precision maps, which affects the accuracy of navigation.
[0004] In a first aspect, the present invention provides a method for determining missing vector features, comprising:
[0005] Obtain the target node in the target image and the display vector elements connected to the target node;
[0006] At least one reference vector feature connected to the target node is obtained based on the standard image;
[0007] The at least one reference vector element and the display vector element are matched, and the missing vector elements of the target image are determined based on the matching result.
[0008] In some embodiments, after acquiring at least one reference vector feature connected to the target node based on a standard image and before matching the at least one reference vector feature with the display vector feature, the method further includes:
[0009] Reference vector elements with target type attributes are selected from the at least one reference vector element, and the displayed vector element is matched based on the selected reference vector elements with target type attributes, wherein the target type attribute and the missing vector element have the same type attribute.
[0010] In some embodiments, acquiring at least one reference vector feature connected to the target node based on a standard image includes:
[0011] At least one target type attribute vector feature is obtained based on the standard image;
[0012] From the at least one target type attribute vector feature, target type attribute vector features connected to the target node are selected to obtain at least one reference vector feature of target type attribute, wherein the target type attribute and the missing vector feature have the same type attribute.
[0013] In some embodiments, matching the at least one reference vector feature and the display vector feature includes:
[0014] Calculate the distance between reference vector features and the displayed vector features that have the same type attributes;
[0015] Based on the distance and the preset distance threshold, determine whether the reference vector element and the corresponding display vector element match;
[0016] or,
[0017] The at least one reference vector feature and the display vector feature are matched based on the hidden Markov model.
[0018] In some embodiments, calculating the distance between reference vector features and display vector features with the same type attribute includes:
[0019] Calculate the Fraser distance or Hausdorff distance between reference vector features and display vector features that have the same type attributes.
[0020] In some embodiments, the reference vector feature includes a plurality of discrete vector sub-features. After acquiring at least one reference vector feature connected to the target node based on a standard image and before matching the at least one reference vector feature with the display vector feature, the method further includes:
[0021] When multiple reference vector elements are obtained, the vector sub-elements of the multiple reference vector elements are connected end to end to obtain the spliced reference vector elements.
[0022] In some embodiments, determining the missing vector features of the target image based on the matching results includes:
[0023] When a match fails, the reference vector feature is identified as the missing vector feature.
[0024] In some embodiments, obtaining the target node in the target image includes:
[0025] Obtain at least one of the following in the target image: a fork node, a merge node, a start node, and an end node.
[0026] In a second aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for determining missing vector elements as described in any of the preceding claims.
[0027] In a third aspect, the present invention provides an electronic device, characterized in that it comprises:
[0028] At least one processor;
[0029] And, a memory communicatively connected to the at least one processor;
[0030] The memory stores a computer program that, when executed by the at least one processor, implements the method for determining missing vector features as described above.
[0031] By employing the above technical solution, the present invention provides a method for determining missing vector elements. This method involves acquiring target nodes in a target image and display vector elements connected to those nodes, obtaining at least one reference vector element connected to the target node based on a standard image, matching the at least one reference vector element with the display vector element, and determining the missing vector elements in the target image based on the matching result. This method utilizes a relatively complete standard map and target image for matching, and determines the missing vector elements in the target image when the reference vector element matching fails. It can automatically and effectively determine missing vector elements, thus avoiding the use of display vector elements with lower actual matching degrees to replace missing vector elements in existing technologies when the target image has missing elements, thereby improving navigation accuracy. Attached Figure Description
[0032] The preferred embodiments of the present invention are described below with reference to the accompanying drawings, in which:
[0033] Figure 1 This is a schematic diagram of the main steps of a method for determining missing vector elements provided in an embodiment of the present invention;
[0034] Figure 2 This is a schematic diagram of a missing ramp on a main road provided by the present invention;
[0035] Figure 3 This is a schematic diagram of a partial missing line on a ramp provided by the present invention;
[0036] Figure 4 This is a schematic flowchart of a method for determining missing vector elements provided in an embodiment of the present invention;
[0037] Figure 5 This is a schematic flowchart of a method for determining missing vector elements provided in another embodiment of the present invention;
[0038] Figure 6 This is a schematic diagram of the spliced reference vector elements provided in an embodiment of the present invention;
[0039] Figure 7 This is a schematic diagram of the electronic device structure provided in an embodiment of the present invention. Detailed Implementation
[0040] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.
[0041] As described in the background section, high-precision maps often contain missing map features, including roads. Roads are linear vector features, and based on different type attributes, they can be categorized as main roads, highways, and ramps. In existing technologies, when performing navigation based on SD-HD route matching, if a route is missing in the HD map, other existing vector features in the HD map with a relatively high matching degree to the SD route are automatically matched as the target output. This fails to automatically and effectively identify the missing vector features, thus affecting navigation accuracy. Therefore, this invention provides a method for determining missing vector features, see [link to relevant documentation]. Figure 1 As shown, Figure 1 This is a flowchart illustrating the main steps of a method for determining missing vector elements provided in an embodiment of the present invention, which may include:
[0042] Step S11: Obtain the target nodes in the target image and the display vector features connected to the target nodes;
[0043] Step S12: Obtain at least one reference vector feature connected to the target node based on the standard image;
[0044] Step S13: Match at least one reference vector feature with the displayed vector feature, and determine the missing vector features of the target image based on the matching result.
[0045] In some embodiments, regions in the target image where missing vector features may exist can be preliminarily determined based on the topological structure of vector features in the target image and the type attributes of the missing vector features to be identified. Taking the type attribute of the missing vector features as an example, such as a ramp... Figure 2 and 3 As shown, Figure 2 This is a schematic diagram of a missing ramp on a main road provided by the present invention. Figure 3 This is a schematic diagram illustrating a partial loss of track on a ramp, as provided by the present invention. Figure 2 and Figure 3 Two types of missing vector features with the attribute "ramp" are provided, from Figure 2 and Figure 3 It is known that when a line suddenly disconnects and a connection termination node appears, there will be missing ramps in the area surrounding the termination node. Similarly, there will also be missing ramps at connection fork nodes, connection merging nodes, and connection starting nodes.
[0046] Based on this, obtaining the target node in the target image in step S11 can specifically be obtaining at least one of the following in the target image: a fork node, a merge node, a start node, and a termination node.
[0047] In some embodiments, obtaining the display vector features connected to the target node in step S11 can be obtaining the display vector features of all types of attributes connected to the target node in the target image; in other embodiments, obtaining the display vector features connected to the target node in step S11 can be obtaining the display vector features of the target type attribute connected to the target node in the target image, wherein the target type attribute and the type attribute of the missing vector feature are the same.
[0048] In some embodiments, step S12 may specifically be:
[0049] Obtain at least one target type attribute vector feature based on a standard image;
[0050] From at least one target type attribute vector feature, select the target type attribute vector features connected to the target node to obtain at least one reference vector feature of the target type attribute.
[0051] The standard image can be more complete than the target image. For example, the standard image can be a standard map, and the target image can be a high-precision map.
[0052] In some embodiments, obtaining at least one reference vector element connected to the target node in step S12 can specifically be: obtaining at least one reference vector element connected to the target node by using navigation route planning in a standard image. Specifically, the target node can be used as the starting point, and multiple points at different directions at a preset distance from the starting point can be used as the endpoints to obtain at least one reference vector element connected to the target node.
[0053] In some embodiments, matching at least one reference vector feature and the display vector feature in step S13 can specifically be as follows:
[0054] Calculate the distance between reference vector features and display vector features that have the same type attributes;
[0055] The system determines whether the reference vector features and the corresponding display vector features match based on the distance and a preset distance threshold.
[0056] In some embodiments, when a reference vector feature of a target type attribute is obtained, calculating the distance between a reference vector feature and a display vector feature with the same type attribute can be done by calculating the distance between a reference vector feature of the target type attribute and a display vector feature of the target type attribute. When multiple types of reference vector features are obtained, after step S12 and before step S13, reference vector features of the target type attribute can be selected from the multiple types of reference vector features to calculate the distance between the reference vector features of the target type attribute and the display vector feature of the target type attribute.
[0057] In other embodiments, when multiple types of reference vector features are obtained, step S13 can also directly match the reference vector features of various types of attributes with the vector features of the same type of attributes, and then determine the missing vector features of the target image based on the reference vector features that failed to match and whose type attribute is the target type attribute.
[0058] In some embodiments, the distance between reference vector features and display vector features with the same type attribute can be calculated as follows:
[0059] Calculate the Fraser distance or Hausdorf distance between reference vector features and display vector features with the same type attributes.
[0060] By calculating the Fraser distance or Hausdorff distance and determining whether the reference vector feature and the corresponding display vector feature match based on the Fraser distance or Hausdorff distance, the accuracy of the distance can be greatly improved, thereby improving the effectiveness of the matching results.
[0061] In some embodiments, determining whether a reference vector feature and a corresponding display vector feature match based on a distance and a preset distance threshold can specifically be as follows:
[0062] When the distance is less than or equal to a preset distance threshold, the reference vector element and the corresponding display vector element are matched. When the distance is greater than the preset distance threshold, the matching fails, the reference vector element does not match the display vector element, and the corresponding reference vector element is missing from the target image. Based on the position information of the reference vector element, the position information of the missing vector element in the target image can be determined. The preset distance threshold can be set according to requirements.
[0063] In other embodiments, matching at least one reference vector feature and the display vector feature in step S13 may specifically be as follows:
[0064] At least one reference vector feature and a displayed vector feature are matched using a Hidden Markov Model (HMM). Matching reference and displayed vector features using an HMM can effectively improve the accuracy of the matching, thus facilitating the accurate identification of missing vector features.
[0065] In this embodiment of the invention, determining the missing vector elements of the target image based on the matching result in step S13 can be: when the matching fails, the reference vector elements are determined as missing vector elements.
[0066] The above describes a method for determining missing vector elements provided by an embodiment of the present invention. This method involves acquiring target nodes and display vector elements connected to the target nodes in a target image, acquiring at least one reference vector element connected to the target nodes based on a standard image, matching the at least one reference vector element with the display vector elements, and determining the missing vector elements in the target image based on the matching result. This method utilizes a relatively complete standard map and target image for matching, and determines the missing vector elements in the target image when the reference vector element matching fails. It can automatically and effectively determine missing vector elements, thus avoiding the use of display vector elements with lower actual matching degrees to replace missing vector elements in existing technologies when the target image has missing elements, which is beneficial for improving navigation accuracy.
[0067] See Figure 4 As shown, Figure 4 This is a flowchart illustrating a method for determining missing vector features according to an embodiment of the present invention, which may include:
[0068] Step S41: Obtain the target nodes in the target image and the display vector elements connected to the target nodes;
[0069] Step S42: Obtain at least one reference vector feature connected to the target node based on the standard image;
[0070] Step S43: Select reference vector features with target type attributes from at least one reference vector feature;
[0071] Step S44: Match the reference vector features and display vector features of the target type attribute, and determine the missing vector features of the target image based on the matching results.
[0072] Step S41 can be implemented in the same way as step S11. For the sake of brevity, it will not be described again here. For details, please refer to the description above.
[0073] In some embodiments, step S42 may specifically be: obtaining at least one reference vector element connected to the target node by using navigation route planning in the standard image. Specifically, the target node may be used as the starting point, and multiple points at different directions at a preset distance from the starting point may be used as the endpoints to use navigation route planning, thereby obtaining at least one reference vector element connected to the target node.
[0074] In some embodiments, step S43 may specifically involve filtering reference vector features from at least one reference vector feature based on the type attribute, which are the same as the missing vector feature to be confirmed, i.e., reference vector features with the target type attribute.
[0075] In some embodiments, step S44 may specifically involve calculating the distance between the reference vector feature of the target type attribute and the display vector feature of the target type attribute; determining whether the reference vector feature of the target type attribute and the corresponding display vector feature match based on the distance and a preset distance threshold; and determining the reference vector feature of the target type attribute as a missing vector feature when the match fails.
[0076] In some embodiments, the distance between the reference vector feature of the target type attribute and the display vector feature of the target type attribute can be calculated as either the Fraser distance or the Hausdorf distance between the reference vector feature of the target type attribute and the display vector feature of the target type attribute.
[0077] By calculating the Fraser distance or Hausdorff distance and determining whether the reference vector features and corresponding display vector features of the target type attribute match based on the Fraser distance or Hausdorff distance, the accuracy of the distance can be greatly improved, thereby improving the effectiveness of the matching results.
[0078] In some embodiments, determining whether the reference vector features and corresponding display vector features of the target type attribute match based on distance and a preset distance threshold can specifically be as follows:
[0079] When the distance is less than or equal to a preset distance threshold, the reference vector feature of the target type attribute and the corresponding display vector feature are matched. When the distance is greater than the preset distance threshold, the matching fails, the reference vector feature of the target type attribute does not match the display vector feature, and the corresponding reference vector feature is missing from the target image. Based on the position information of the reference vector feature of the target type attribute, the position information of the missing vector feature in the target image can be determined. The preset distance threshold can be set according to requirements.
[0080] In other embodiments, step S44 may specifically involve matching the reference vector features of the target type attribute with the displayed vector features of the target type attribute based on a Hidden Markov Model (HMM); when the matching fails, the reference vector features of the target type attribute are identified as missing vector features. Matching the reference vector features of the target type attribute with the displayed vector features of the target type attribute based on an HMM can effectively improve the accuracy of the matching, thereby facilitating the accurate identification of missing vector features.
[0081] The above describes another embodiment of the method for determining missing vector elements provided by the present invention. By acquiring target nodes and display vector elements connected to the target nodes in the target image, and obtaining at least one reference vector element connected to the target nodes based on a standard image, reference vector elements with target type attributes are selected from the at least one reference vector element. This allows for direct matching of the reference vector elements with target type attributes and the corresponding display vector elements, reducing the computational resources required for the matching process and facilitating the rapid and convenient determination of missing vector elements in the target image based on the matching results. Furthermore, it avoids the prior art's practice of replacing missing vector elements with display vector elements that have a lower actual matching degree when the target image is missing, thus improving navigation accuracy.
[0082] In other embodiments, the reference vector features may include multiple discrete vector sub-features. When multiple reference vector features are obtained, the vector sub-features of the multiple reference vector features can be concatenated end-to-end to obtain the final number of reference vector features. See the description in the following embodiments for details. It should be noted that this embodiment can be based on the above... Figure 1 or Figure 4 The corresponding implementation example is based on... Figure 4 The corresponding implementation examples are described below.
[0083] See Figure 5 The above, Figure 5 This is a flowchart illustrating a method for determining missing vector features according to another embodiment of the present invention, which may include:
[0084] Step S51: Obtain the target nodes in the target image and the display vector elements connected to the target nodes;
[0085] Step S52: Obtain at least one reference vector feature connected to the target node based on the standard image;
[0086] Step S53: Select reference vector features with target type attributes from at least one reference vector feature;
[0087] Step S54: When multiple target type attributes are obtained, the vector sub-features of the multiple target type attributes are connected end to end to obtain the spliced reference vector features.
[0088] Step S55: Match the stitched reference vector elements and the displayed vector elements, and determine the missing vector elements of the target image based on the matching results.
[0089] Steps S51-S53 and S55 can be implemented using the same method as steps S41-S44. It should be noted that in step S55, the spliced reference vector features are the reference vector features of the target type attribute obtained after splicing. The final number can be greater than or equal to the number of reference vector features before splicing. When executing step S55, it can be adaptively adjusted based on the method provided in step S44.
[0090] As an example, see Figure 6 As shown, Figure 6 This is a schematic diagram of the stitched reference vector elements provided in an embodiment of the present invention. AmnC and BmnD are reference vector elements for two target type attributes obtained before stitching. By connecting the two target type attribute vector sub-elements end-to-end, four reference vector elements—AmnC, BmnD, AmnD, and BmnC—are obtained, serving as the final reference vector elements for the target type attributes.
[0091] The above describes another embodiment of the method for determining missing vector elements, which achieves the same beneficial effects as any of the above embodiments. Furthermore, by concatenating the vector sub-elements of multiple reference vector elements end-to-end, a spliced reference vector element can be obtained, ensuring a complete reference vector element. Additionally, concatenating the vector sub-elements end-to-end avoids situations where a vector sub-element is divided into two segments by a forking or merging node, and the two segments correspond to the same identifier. In such cases, the identifier might be duplicated, leading to the deletion of one identifier and its corresponding vector sub-element portion, thus preventing incomplete reference vector elements.
[0092] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.
[0093] In another aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for determining missing vector elements as described in any of the above embodiments. This computer-readable storage medium may be a storage device comprising various electronic devices; optionally, in embodiments of the present invention, the computer-readable storage medium is a non-transitory computer-readable storage medium.
[0094] Another aspect of the present invention provides an electronic device that may include at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program that, when executed by the at least one processor, implements the method for determining missing vector elements as described in any of the above embodiments.
[0095] See Figure 7 As shown, Figure 7 The example shows a configuration where memory 71 and processor 72 are connected via a bus, and both memory 71 and processor 72 are configured with only one instance.
[0096] In other embodiments, the electronic device may include multiple memories 71 and multiple processors 72. The program executing the method for determining missing vector elements in any of the above embodiments may be divided into multiple subroutines, each of which may be loaded and run by a processor 72 to perform different steps of the method for determining missing vector elements in the above embodiments. Specifically, each subroutine may be stored in a different memory 71, and each processor 72 may be configured to execute programs in one or more memories 71 to jointly implement the method for determining missing vector elements in the above embodiments.
[0097] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.
Claims
1. A method for determining missing vector features, characterized in that, The vector features include road features, and the method includes: Obtain target nodes in a target image and display vector features connected to the target nodes, wherein the target image is a map to be detected; At least one reference vector feature connected to the target node is obtained based on a standard image, wherein the standard image is a standard map; The at least one reference vector element and the display vector element are matched, and the missing vector elements of the target image are determined based on the matching result.
2. The method according to claim 1, characterized in that, After acquiring at least one reference vector feature connected to the target node based on a standard image and before matching the at least one reference vector feature with the display vector feature, the method further includes: Reference vector elements with target type attributes are selected from the at least one reference vector element, and the displayed vector element is matched based on the selected reference vector elements with target type attributes, wherein the target type attribute and the missing vector element have the same type attribute.
3. The method according to claim 1, characterized in that, The step of acquiring at least one reference vector feature connected to the target node based on a standard image includes: At least one target type attribute vector feature is obtained based on the standard image; From the at least one target type attribute vector feature, target type attribute vector features connected to the target node are selected to obtain at least one reference vector feature of target type attribute, wherein the target type attribute and the missing vector feature have the same type attribute.
4. The method according to any one of claims 1 to 3, characterized in that, The step of matching the at least one reference vector feature with the display vector feature includes: Calculate the distance between reference vector features and the displayed vector features that have the same type attributes; Based on the distance and the preset distance threshold, determine whether the reference vector element and the corresponding display vector element match; or, The at least one reference vector feature and the display vector feature are matched based on the hidden Markov model.
5. The method according to claim 4, characterized in that, The distance between reference vector features and displayed vector features with the same type attributes is calculated as follows: Calculate the Fraser distance or Hausdorff distance between reference vector features and display vector features that have the same type attributes.
6. The method according to any one of claims 1 to 3, characterized in that, The reference vector feature includes multiple discrete vector sub-features. After obtaining at least one reference vector feature connected to the target node based on the standard image and before matching the at least one reference vector feature with the display vector feature, the method further includes: When multiple reference vector elements are obtained, the vector sub-elements of the multiple reference vector elements are connected end to end to obtain the spliced reference vector elements.
7. The method according to claim 1, characterized in that, The step of determining the missing vector features of the target image based on the matching results includes: When a match fails, the reference vector feature is identified as the missing vector feature.
8. The method according to claim 1, characterized in that, The acquisition of target nodes in the target image includes: Obtain at least one of the following in the target image: a fork node, a merge node, a start node, and an end node.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method for determining missing vector elements as described in any one of claims 1 to 8.
10. An electronic device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that, when executed by the at least one processor, implements the method for determining missing vector features as described in any one of claims 1 to 8.