Method for comparing map sections of a digital road map
Patent Information
- Application Number
- DE102023209267
- Authority / Receiving Office
- DE · DE
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-09-22
- Publication Date
- 2025-03-27
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
[0001] The invention relates to a method for comparing a first map section of a digital road map with a second map section of the digital road map that at least partially overlaps the first map section, a device, a computer program and a machine-readable storage medium. State of the art
[0002] The published patent application CN 1 13 192 142 A discloses a map construction method.
[0003] The published patent application CN 1 13 066 105 A discloses a positioning and mapping system.
[0004] The published patent application CN 1 15 077 519 A discloses a positioning and mapping method.
[0005] In simultaneous localization and mapping (SLAM), sensor data must be regularly compared and merged with map data from different journeys. Road markings play a crucial role, especially in automotive-related camera-based SLAM systems. Due to their good visibility and widespread availability, they represent a fundamental element in HD mapping. At the same time, a vehicle can locate itself relatively precisely, especially laterally, using the road markings. To create accurate maps, it is often not enough to use just one journey along a route; instead, maps are created using an entire fleet of vehicles, so-called crowdsourced maps. For this purpose, the individual maps created by one vehicle are aligned and merged with other maps created by other vehicles.A fundamental problem in mapping and localization, however, is often that only local sections of the map are used, yet only one road marking is detected in many road sections. This can occur due to perception errors. It can also be the case that only one road marking exists in a road section for an extended period of time; for example, the right lane marking is the only road marking on some roads. Since many algorithms in HD (high definition) road mapping are based on calculating relative rotations and translations between two map sections, the situation in which only one tram marking exists or can be matched is often a critical case. In this case, a clear roll angle cannot be determined. Disclosure of the invention
[0006] The object underlying the invention is to provide a concept for comparing a first map section of a digital road map with a second map section of the digital road map that at least partially overlaps the first map section.
[0007] This object is achieved by means of the respective subject matter of the independent claims. Advantageous embodiments of the invention are the subject matter of the respective dependent subclaims.
[0008] According to a first aspect, a method is provided for comparing a first map section of a digital road map with a second map section of the digital road map that at least partially overlaps the first map section, comprising the following steps: determining that a first relative rotation between the two map sections cannot be clearly determined, determining a second relative rotation between a third map section of the digital road map and a fourth map section of the digital road map that at least partially overlaps the third map section, wherein the third map section and the fourth map section are adjacent to the first map section and the second map section,
[0009] Matching the first map section with the second map section, wherein the matching comprises determining a relative rotation between the first map section and the second map section based on the determined second relative rotation.
[0010] According to a second aspect, a device is provided which is configured to carry out all the steps of the method according to the first aspect.
[0011] According to a third aspect, a computer program is provided, comprising instructions which, when the computer program is executed by a computer, for example by the device according to the second aspect, cause the computer to carry out a method according to the first aspect.
[0012] According to a fourth aspect, a machine-readable storage medium is provided on which the computer program according to the third aspect is stored.
[0013] The invention is based on and includes the finding that the above object is achieved by using information, in this case the second relative rotation, from adjacent map sections of the two map sections to be compared in the event of a degenerate situation, i.e. when a first relative rotation between the two map sections cannot be clearly determined. The determined second relative rotation between the third map section and the fourth map section is thus used according to the concept described here to compare the first map section with the second map section. The first relative rotation, which could not previously be clearly determined, is now determined according to the concept described here using the determined second relative rotation.
[0014] Thus, even in the case of a degenerate situation, the first map section can be compared with the second map section.
[0015] A degenerate situation is a situation in which it is determined or specified that a first relative rotation between the two map sections cannot be unambiguously determined. This is the case, for example, if only one road marker exists or is detected in one of the map sections, whereas multiple road markers exist in the other map section. This is the case, for example, if only one road marker exists or is detected in each of the map sections.
[0016] If the relative rotation between the two map sections to be compared cannot be clearly determined, a so-called roll angle ambiguity exists.
[0017] To resolve such a roll angle ambiguity, according to the concept described here, additional information is used from adjacent map sections: the second relative rotation between the two adjacent map sections, i.e. the roll angle between the two adjacent map sections.
[0018] For example, it is planned that the roll angle, i.e. the relative rotation, between the two adjacent map sections is propagated into the two map sections to be aligned.
[0019] Thus, according to one embodiment of the method, determining the relative rotation between the first map section and the second map section comprises propagating the determined second relative rotation into the first and second map sections.
[0020] This provides the technical advantage, for example, that the relative rotation between the first map section and the second map section can be determined efficiently.
[0021] In one embodiment of the method, it is provided that a relative delta pose between the first map section and the second map section is determined, wherein a set of inliers is determined based on the determined delta pose, wherein a principal axis transformation of the determined set of inliers is carried out, wherein based on the principal axis transformation it is determined that a first relative rotation between the two map sections cannot be uniquely determined.
[0022] This provides, for example, the technical advantage that the presence of a degenerate situation can be efficiently detected or determined.
[0023] In one embodiment of the method, it is provided that the principal axis transformation comprises performing a singular value decomposition of the set of inliers in order to determine the principal axes, wherein it is determined based on the singular value decomposition that a first relative rotation between the two map sections cannot be uniquely determined.
[0024] This provides, for example, the technical advantage of being able to efficiently determine when a degenerate situation exists.
[0025] In one embodiment of the method, if two of three singular values determined according to the singular value decomposition are less than or equal to a predetermined limit, it is determined that all inliers lie on a main axis, wherein it is determined that a first relative rotation between the two map sections cannot be uniquely determined.
[0026] This provides, for example, the technical advantage that a degenerate situation can be efficiently detected.
[0027] In one embodiment of the method, it is provided that the set of inliers is determined using a nearest neighbor heuristic using a distance bound by specifying that each point of a lane marking represented by the first map section is an inlier if there is a point in the second map section that is located within the distance bound to the corresponding point of the lane marking represented by the second map section.
[0028] This provides, for example, the technical advantage that the number of inliers can be determined efficiently.
[0029] Device features result analogously from corresponding process features and vice versa.
[0030] The method is carried out, for example, by means of the device.
[0031] For example, the method is a computer-implemented method.
[0032] For example, a comparison between two map sections involves determining a relative delta pose between these two map sections.
[0033] A third and / or a fourth map section within the meaning of the description is, for example, an immediately adjacent map section in relation to the first map section and / or in relation to the second map section.
[0034] For example, the relative rotation between the first map section and the second map section is specified to be equal to the second relative rotation between the third map section and the fourth map section.
[0035] A relative rotation can also be referred to as a roll angle.
[0036] For example, determining a relative delta pose involves determining a relative rotation and a relative translation.
[0037] A map section within the meaning of the description was or is determined, for example, using environmental sensor data from one or more environmental sensors of a motor vehicle.
[0038] An environmental sensor within the meaning of the description is, for example, one of the following environmental sensors: radar sensor, LiDAR sensor, image sensor, in particular image sensor of a video camera, ultrasonic sensor, infrared sensor and magnetic field sensor.
[0039] Thus, a map section within the meaning of the description can be generated or created, for example, using LiDAR data and / or radar data and / or video camera data and / or infrared data and / or magnetic field data and / or ultrasound data.
[0040] Environment sensor data as defined in the description describes the environment of the motor vehicle.
[0041] A road marking within the meaning of the description is, for example, one of the following road markings: longitudinal marking, area marking, boundary marking, transverse marking, parking area marking.
[0042] A road marking can also be called a lane marking.
[0043] A longitudinal marking includes, for example, one of the following longitudinal markings: guide line, warning line, lane marking, lane marking, single lane marking, exit line, double lane marking.
[0044] A longitudinal marking may, for example, comprise: a continuous line, a broken line, a double line which is both continuous and broken, a continuous double line, a broken double line.
[0045] The embodiments and forms of embodiment described here can be combined with each other in any way, even if this is not explicitly described.
[0046] For example, the device is programmed to execute the computer program.
[0047] The invention is explained in more detail below using preferred embodiments. These show: Fig. 1 a flowchart of a method according to the first aspect, Fig. 2 a device according to the second aspect, Fig. 3 a machine-readable storage medium according to the fourth aspect, Fig. 4 two map sections of a digital road map before comparison, Fig. 5 the two map sections of the Fig. 4 after a comparison, Fig. 6 two map sections of a digital road map before comparison, Fig. 7 the two map sections of the Fig. 6 after a comparison, Fig. 8 two map sections of a digital road map before comparison.
[0048] In the following, the same reference symbols may be used for the same features.
[0049] Fig. 1 shows a flowchart of a method for comparing a first map section of a digital road map with a second map section of the digital road map that at least partially overlaps the first map section, comprising the following steps: Determine 101 that a first relative rotation between the two map sections cannot be uniquely determined, Determining 103 a second relative rotation between a third map section of the digital road map and a fourth map section of the digital road map at least partially overlapping the third map section, wherein the third map section and the fourth map section are adjacent to the first map section and the second map section, Matching 105 the first map section with the second map section, wherein the matching comprises determining 107 a relative rotation between the first map section and the second map section based on the determined second relative rotation.
[0050] Fig. 2 shows a device 201 which is configured to carry out all steps of the method according to the first aspect.
[0051] Fig. 3 shows a machine-readable storage medium 301 on which a computer program 303 is stored. The computer program 303 includes instructions that, when executed by a computer, cause the computer program 303 to execute a method according to the first aspect.
[0052] Fig. Figure 4 shows two map sections of a digital road map that are to be compared. Fig. 4 a first map section 401 and a second map section 403.
[0053] The first map section 401 includes a first street 405. The second map section 403 includes a second street 407.
[0054] The first road, 405, has two solid lines, 409 and 411, as lane markings. A dashed line, 413, also runs as the center line between the two lane markings, 409 and 411.
[0055] For the second map section 403, three solid lines 415, 417, 419 were recognized, as well as a dashed line 421 and a short line 423, which appears to belong to the dashed line 421.
[0056] Here, two solid lines 415, 417 are very close to each other.
[0057] The solid lines 415, 417, and 419 form the boundary lines or lane boundaries of road 407. The dashed line 421 is a median. The short line 423, for example, is an error detection.
[0058] Using well-known alignment algorithms, the two map sections can be aligned. This is shown, for example, by the Fig. 5.
[0059] Fig. 6 shows a first map section 601 and a second map section 603 which are to be compared.
[0060] In contrast to the first map section 401 according to Fig. 4 is with regard to the first map section 601 of the Fig. 6 only the dashed line 413 exists to align the two map sections. After aligning the two map sections, which is exemplified in Fig. 7, the roll angle cannot be determined clearly. The calibration has proven to be degenerate. Therefore, a degenerate situation exists. The roll angle is in Fig. 7 is exemplified by a circular arrow with the reference number 701.
[0061] Fig. 8 shows a first map section 801 and a second map section 803, which at least partially overlap so that they can be compared based on their overlap.
[0062] According to the concept described here, in the case of a degenerate situation, which is Fig. 6 and Fig.7 is shown as an example, it is intended that information from adjacent map sections is used: the roll angle between the two adjacent map sections.
[0063] Thus, the concept described here is based in particular on the recognition or detection of a degenerate situation.
[0064] As part of the comparison, for example, lane markings of the map sections are associated.
[0065] The concept described here thus detects degenerate lane associations, in particular line associations, whereby the geometric underdetermination of the problem, i.e. the association, is resolved with the help of additional information.
[0066] An association can also be called a “matching”.
[0067] For example, two map sections are compared by an algorithm, whereby the respective lane markings, especially lines, are associated. Such algorithms are known per se. In particular, the comparison result is used to determine whether a degenerate situation exists or not. If such a degenerate situation exists, the associated tram markings essentially all lie on a straight line. For example, a principal axis analysis can be used to determine whether a point cloud is very similar to a straight line.
[0068] To resolve a roll angle ambiguity, for example, additional information from two adjacent map sections that at least partially overlap is used. For example, these two adjacent map sections are compared, with the respective lane markings being associated during the comparison. If no degenerate situation exists, a correspondingly determined roll angle between these two adjacent map sections can be used to determine the roll angle between the map sections originally to be compared.
[0069] For example, the inliers are determined based on the relative delta pose between the two map sections to be compared. Inliers are determined, for example, using a nearest neighbor heuristic, in particular using a distance bound. This means that every point (of a line) from the first map section counts as an inlier if there is a point in the second map section that is close enough to the point from the second map. Now, for example, the set of all inliers is considered. Based on the set of all inliers, the principal axes are calculated. The principal axes can be estimated, for example, using a singular value decomposition. If, for example, two of the three singular values are very small, then all inliers lie on one axis and this is a degenerate situation.
[0070] If a degenerate situation is detected, it is remedied by adding additional information or information, in particular the second roll angle. For this purpose, additional neighboring map sections are used. If no degenerate situation is detected for these neighboring map sections as part of a comparison of these two map sections, the roll angle between the two neighboring map sections can be propagated into the map section pair with the degenerate comparison, for example. To enable the roll angle to be propagated further, it is provided, for example, that the neighboring map sections lie within the same digital road map. For example, the neighboring map sections within the same road map with a sufficient overlap are used.If the digital road map was or is generated using additional information, such as inertial sensors, odometry, etc., this additional information can also be used, for example, to determine the relative rotation between the first and the second map section, i.e., to propagate the second roll angle.
[0071] The concept described here is not necessarily focused on lines; point-like features such as posts or planes, etc., can also be used for alignment. A mixed use of geometry types is envisaged, for example.
[0072] The map sections can, for example, be generated from camera data and can, for example, be generated in addition to or instead of data from other environmental sensors, for example LiDAR sensors or radar sensors and / or other environmental sensors as already described above. QUOTES CONTAINED IN THE DESCRIPTION
[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature
[0000] CN 1 13 192 142 A
[0002] CN 1 13 066 105 A
[0003] CN 1 15 077 519 A
[0004]
Claims
[1] Method for comparing a first map section (401, 601, 801) of a digital road map with a second map section (403, 603, 803) of the digital road map at least partially overlapping the first map section (401, 601, 801), comprising the following steps: Determining (101) that a first relative rotation between the two map sections cannot be uniquely determined, Determining (103) a second relative rotation between a third map section of the digital road map and a fourth map section of the digital road map at least partially overlapping the third map section, wherein the third map section and the fourth map section are adjacent to the first map section (401, 601, 801) and the second map section (403, 603, 803), Comparing (105) the first map section (401, 601, 801) with the second map section (403, 603, 803), wherein the comparability comprises determining (107) a relative rotation between the first map section (401, 601, 801) and the second map section (403, 603, 803) based on the determined second relative rotation. [2] The method of claim 1, wherein determining the relative rotation between the first map section (401, 601, 801) and the second map section (403, 603, 803) comprises propagating the determined second relative rotation into the first (401, 601, 801) and the second map section (403, 603, 803). [3] Method according to claim 1 or 2, wherein a relative delta pose between the first map section (401, 601, 801) and the second map section (403, 603, 803) is determined, wherein based on the determined delta pose a set of inliers is determined, wherein a principal axis transformation of the determined set of inliers is performed, wherein based on the principal axis transformation it is determined that a first relative rotation between the two map sections cannot be uniquely determined. [4] The method of claim 3, wherein the principal axis transformation comprises performing a singular value decomposition of the set of inliers to determine the principal axes, wherein it is determined based on the singular value decomposition that a first relative rotation between the two map sections cannot be uniquely determined. [5] The method of claim 4, wherein, if two of three singular values determined according to the singular value decomposition are less than or equal to a predetermined bound, it is determined that all inliers lie on a main axis, wherein it is determined that a first relative rotation between the two map sections cannot be uniquely determined. [6] Method according to one of claims 3 to 5, wherein the set of inliers is determined using a nearest neighbor heuristic using a distance bound by specifying that each point of a lane marking (409, 411, 413) represented by the first map section is an inlier if there is a point in the second map section (403, 603, 803) that is within the distance bound to the corresponding point of the lane marking (415, 417, 419, 421, 423) represented by the second map section. [7] Device (201) which is arranged to carry out all the steps of the method according to one of the preceding claims. [8] Computer program (303) comprising instructions which, when the computer program (303) is executed by a computer, cause the computer to carry out a method according to one of claims 1 to 6. [9] Machine-readable storage medium (301) on which the computer program (303) according to claim 8 is stored.
Citation Information
Patent Citations
Positioning and mapping method and system based on fusion of laser radar and inertial measurement unit
CN113066105A
High-precision map construction method and device in complex environment and computer equipment
CN113192142A
Positioning mapping method and device based on template matching and laser inertial navigation loose coupling
CN115077519A
Method for positioning a map representation of a vehicle's surroundings in a semantic road map
DE102021209786A1
Method and arrangement for sourcing of location information, generating and updating maps representing the location
EP3617749B1