Element updating method, device, computer equipment and storage medium

Abstract

The present application relates to an element updating method, device, computer equipment and storage medium. The method relates to the field of intelligent driving technology, including: if a first key frame in a historical trajectory has a loop relationship with a second key frame in a current trajectory, point cloud alignment is performed on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine the alignment transformation relationship, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory; based on the alignment transformation relationship, the element position information bound to the target key frame is updated, and the target key frame includes: the first key frame, and / or, the second key frame. The method is used to improve the efficiency of reconstructing elements in a map.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to an element updating method, device, computer equipment and storage medium. Background Art

[0002] At present, the mainstream method for building urban scene maps is crowdsourcing mapping. Crowdsourcing mapping solutions are generally divided into two main modules: vehicle-side and cloud-side. The vehicle-side is equipped with surveying and mapping equipment, and performs some lightweight data processing in real time. A large amount of processed data will be transmitted to the cloud, which is responsible for fusing the large amount of fragmented data sent back to reconstruct the elements in the map. This method involves uploading and sending data, and the large amount of fragmented data uploaded by the vehicle side needs to be processed by the cloud through complex aggregation algorithms to reconstruct the elements in the map. Therefore, the efficiency is often low when reconstructing the elements in the map, and it is impossible to generate an instantly available reconstructed map based on the data on the vehicle side in real time. Summary of the invention

[0003] Based on this, it is necessary to provide an element updating method, apparatus, computer device, computer readable storage medium and computer program product to address the above technical issues.

[0004] In a first aspect, the present application provides an element updating method, comprising:

[0005] Obtaining a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory;

[0006] Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship;

[0007] Based on the alignment transformation relationship, the element position information bound to the target key frame is updated, and the target key frame includes: the first key frame and / or the second key frame.

[0008] In one embodiment, updating the element position information bound to the target keyframe based on the alignment transformation relationship includes:

[0009] The pose of the target key frame is adjusted based on the alignment transformation relationship to align the historical trajectory with the current trajectory; and the element position information bound to the target key frame is updated according to the adjusted pose of the target key frame.

[0010] In one embodiment, adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory includes:

[0011] Based on the alignment transformation relationship and the target constraint condition, adjusting the pose of the target key frame to align the historical trajectory with the current trajectory;

[0012] The target constraint condition includes at least: a constraint condition on the relative posture between the first key frame and the second key frame.

[0013] In one embodiment, the target constraint condition further includes:

[0014] Constraints on the recursive pose between the first key frame and adjacent key frames;

[0015] and / or,

[0016] An error constraint condition between the position information of the first key frame and the corresponding positioning information of the first key frame.

[0017] In one of the embodiments, after updating the element position information bound to the target key frame according to the adjusted position of the target key frame, the method further includes:

[0018] Acquire the aligned historical trajectory and the aligned current trajectory;

[0019] The element position information bound to the aligned historical track is merged with the element position information bound to the aligned current track to obtain the element position information bound to the target track.

[0020] In one embodiment, the method further comprises:

[0021] Acquire perception information collected by the vehicle on a target trajectory, wherein the target trajectory includes: the historical trajectory, or the current trajectory;

[0022] Perform element tracking and matching based on the perception information to determine multiple associated position information of the elements;

[0023] Based on the multiple associated position information of the elements, the element position information bound to the target key frame is reconstructed.

[0024] In one embodiment, reconstructing the element position information bound to the first key frame and the second key frame based on the plurality of associated position information of the elements includes:

[0025] Determine, from the plurality of associated position information of the element, a plurality of position information associated with each shape point in the element;

[0026] According to a plurality of position information associated with each shape point in the element, the position information of each shape point of the element is determined to reconstruct the element position information.

[0027] In one embodiment, determining the position information of each shape point of the element according to the plurality of position information associated with each shape point in the element to reconstruct the element position information includes:

[0028] Determine the weight of the plurality of position information corresponding to each shape point;

[0029] For each shape point in the element, weighted sum is performed on the multiple position information corresponding to the multiple position information according to the weight of the multiple position information to determine the position information of each shape point of the element, so as to reconstruct the element position information;

[0030] The weight of each piece of location information in the plurality of pieces of location information is determined according to at least one of the following information:

[0031] a lateral distance between each piece of position information and the vehicle;

[0032] a longitudinal distance between each piece of position information and the vehicle;

[0033] The observation angle of the vehicle corresponding to each piece of position information.

[0034] In one embodiment, determining the position information of each shape point of the element according to the plurality of position information associated with each shape point in the element to reconstruct the element position information includes:

[0035] Determine the position information of each shape point of the element according to the plurality of position information associated with each shape point in the element;

[0036] The position information of the plurality of shape points of the element is fitted to reconstruct the position information of the element.

[0037] In a second aspect, the present application also provides an element updating device, including:

[0038] A determination module is used to obtain a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory; perform loop detection on the first key frame and the second key frame, and when a loop relationship is detected between the first key frame and the second key frame, perform point cloud alignment on point cloud data corresponding to the first key frame and point cloud data corresponding to the second key frame to determine an alignment transformation relationship;

[0039] An updating module is used to update the element position information bound to the target key frame based on the alignment transformation relationship, and the target key frame includes: the first key frame and / or the second key frame.

[0040] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0041] Obtaining a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory;

[0042] Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship;

[0043] Based on the alignment transformation relationship, the element position information bound to the target key frame is updated, and the target key frame includes: the first key frame and / or the second key frame.

[0044] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0045] Obtaining a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory;

[0046] Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship;

[0047] Based on the alignment transformation relationship, the element position information bound to the target key frame is updated, and the target key frame includes: the first key frame and / or the second key frame.

[0048] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0049] Obtaining a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory;

[0050] Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship;

[0051] Based on the alignment transformation relationship, the element position information bound to the target key frame is updated, and the target key frame includes: the first key frame and / or the second key frame.

[0052] In the above-mentioned feature updating method, when there is a loop relationship between the first key frame in the current trajectory of the vehicle and the second key frame in the historical trajectory, it can be determined that the same route has been passed. At this time, according to the point cloud data corresponding to the recorded first key frame and the second key frame, the feature position information bound to the first key frame and / or the second key frame can be updated. In this way, when passing the same route again, the elements in the map can be updated, so that accurate feature information can be reconstructed in real time during driving on the same route again, generating an instantly available reconstructed map, and improving the efficiency of reconstructing elements in the map. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 A schematic diagram of the process of updating a feature Figure 1 ;

[0055] Figure 2A A schematic diagram of a posture;

[0056] Figure 2B A detailed flowchart of an element updating process;

[0057] Figure 3A FIG2 is a flowchart diagram of a method for updating elements;

[0058] Figure 3B A schematic diagram of a process of lane reconstruction;

[0059] Figure 4A schematic diagram of calculating the distance between lane lines;

[0060] Figure 5 It is a schematic diagram of a shape point aggregation;

[0061] Figure 6 Schematic diagram of a process of updating an element three;

[0062] Figure 7 It is a flowchart of an incremental self-learning mapping system;

[0063] Figure 8 It is a structural block diagram of a factor updating device;

[0064] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] High-precision maps are extremely accurate and can achieve lane-level positioning. They are an indispensable key information source for human-like intelligent driving of vehicles. However, the traditional high-precision map production process is cumbersome, the cycle is long, and it requires a lot of manpower and material resources. Its production must first be configured with high-performance sensors to collect vehicle information. After professionals collect surveying and mapping data, they must also perform algorithm processing, manual labeling, manual quality inspection and correction. Once the road changes, it is necessary to re-plan the collection data and update it, and the response speed is slow. For this reason, the mapping scope of high-precision maps is usually simple and difficult to change scenes such as highways.

[0067] With the rapid development of intelligent driving, people's use of the Navigate on Autopilot (NOA) function is no longer limited to high-speed scenarios, and the demand for urban navigation assisted driving is increasing. However, urban roads are wide, the scenarios are diverse, and they change rapidly. The difficulty of implementing traditional mapping solutions in urban areas is exponentially increasing. At present, the mainstream method for building urban scene maps is crowdsourcing mapping. This solution is mainly divided into two modules: the vehicle side and the cloud side. The vehicle is equipped with low-cost mapping equipment. The vehicle side performs lightweight data processing in real time. The processed data is uploaded to the cloud side, and the cloud side performs fusion processing on a large amount of fragmented data. With the continuous increase of observation data, the uncertainty of the scene gradually decreases, and the integrity and accuracy of the map are improved. When the integrity and accuracy of the map constructed by crowdsourcing data meet the requirements, it can be released to users for download and use. The crowdsourcing mapping solution involves uploading and sending data, and the large amount of fragmented data uploaded by the vehicle side needs to be processed by the cloud side through complex aggregation algorithms to reconstruct the elements in the map. Therefore, it is often inefficient when reconstructing the elements in the map, and it is impossible to generate an immediately available reconstructed map based on the data on the vehicle side in real time.

[0068] The element updating method provided in the embodiment of the present application can update the elements in the map when passing the same route again, so that accurate element information can be reconstructed in real time during driving on the same route again, generating an immediately available reconstructed map, thereby improving the efficiency of reconstructing the elements in the map.

[0069] The factor updating method provided in the embodiment of the present application proposes an online incremental self-map solution for the user's daily commuting application scenario. The map is incrementally constructed by self-learning on the vehicle side without manual intervention, and data upload between the vehicle side and the cloud is also avoided. When navigating by self-driving, the vehicle-side user can choose to memorize the route for subsequent autonomous driving. After the system enters the automatic learning mode, the user can terminate the route learning at any time, and the map constructed during the learning process will be saved after the end. When passing the route again in the future, it will automatically enter the learning state again to make up for the unreasonable or incomplete map constructed before due to occlusion, change, etc. In this way, when the user uses the route for assisted driving, the map will become more and more complete, the effect of autonomous driving will become closer to that of humans, and the performance of the vehicle will become more and more intelligent.

[0070] In this application, the mapping process involves processes such as feature reconstruction, feature update, feature matching and fusion, topology post-processing, and layer quality inspection.

[0071] In an exemplary embodiment, Figure 1 A schematic diagram of the process of updating a feature Figure 1The method involves the element updating process in the mapping process, and the method includes the following steps S100 to S103:

[0072] S100: Obtain a first key frame in a historical trajectory and a second key frame in a current trajectory.

[0073] The current track is the current vehicle driving track, and the historical track can be any one or more vehicle driving tracks that have been saved.

[0074] For example, for a commuting scenario, the current trajectory may be the commuting route that the vehicle traveled today, and the historical trajectory may be the commuting route that the vehicle traveled yesterday.

[0075] The first key frame and the second key frame are bound with element position information corresponding to the trajectory.

[0076] S101: Perform loop detection on the first key frame and the second key frame to detect that a loop relationship exists between a first key frame in a historical trajectory and a second key frame in a current trajectory.

[0077] Before the above step S101, key frames in the historical trajectory and key frames in the current trajectory may be saved.

[0078] For example, key frames may be selected from the trajectory at a certain distance, and the element position information corresponding to the trajectory within the certain distance may be bound to the selected key frames.

[0079] For example, assume that the key frame pose matrix is ,in, Represents the vehicle direction, Represents the vehicle position; feature point The representation after switching to the center of the keyframe is: .

[0080] The advantage of selecting a key frame in the above manner is that after the trajectory is aligned, only the pose at the key frame needs to be updated to achieve the purpose of updating the position of all elements.

[0081] It should be noted that the first key frame may refer to one or more key frames, and the second key frame may also refer to one or more key frames.

[0082] If the user drives along a previously remembered route again, the user can preliminarily screen the areas in the historical trajectory and the current trajectory that may have a loop relationship based on the geographic location. If the current trajectory only partially overlaps with the historical trajectory, only the overlapping area will be updated later, and the overlapping route data will be stored in the database.

[0083] After determining that there is an overlapping area between the historical trajectory and the current trajectory, the driving direction can be calculated based on the trajectory to determine candidate loop pairs (such as the first key frame and the second key frame mentioned above) with the same driving direction in the overlapping area.

[0084] When it is determined that a first key frame in the historical trajectory and a second key frame in the current trajectory have a loop relationship, the first key frame and the second key frame are candidate loop pairs.

[0085] In some embodiments, when the position difference between a first key frame in the historical trajectory and a second key frame in the current trajectory is less than a certain threshold, the first key frame and the second key frame are determined to be a candidate loop pair.

[0086] In an embodiment of the present application, it can be determined whether a first key frame in the historical trajectory has a loop relationship with a second key frame in the current trajectory. When it is determined that a first key frame in the historical trajectory has a loop relationship with a second key frame in the current trajectory, the following step S102 is performed.

[0087] S102: performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship.

[0088] If both the first key frame and the second key frame are multiple key frames, the laser point cloud data and vector data corresponding to each key frame included in the first key frame can be obtained respectively, and the data of the multiple key frames included in the first key frame can be spliced ​​to obtain a local map within a certain distance. For example, a local map within 50 meters can be spliced.

[0089] Exemplarily, the point cloud data corresponding to the first key frame may be a source point cloud set. Taking the first key frame including n key frames as an example, it may be expressed as: ,in, Represents a key frame The corresponding point cloud data, Represents a key frame The corresponding point cloud data.

[0090] Exemplarily, the point cloud data corresponding to the second key frame may be a target point cloud set. Taking the second key frame including n key frames as an example, it may be expressed as: ,in, Represents a key frame The corresponding point cloud data, Represents a key frame The corresponding point cloud data.

[0091] In some embodiments, the normal distribution transform (NDT) or iterative closest point (ICP) algorithm can be used to as well as Perform point cloud alignment to determine the transformation relationship between the two, that is, the alignment transformation relationship mentioned above .

[0092] S103: Based on the alignment transformation relationship, update the element position information bound to the target key frame.

[0093] The target key frame includes: a first key frame, and / or a second key frame.

[0094] In some embodiments, the element position information bound to the first key frame may be updated based on the alignment transformation relationship.

[0095] In some embodiments, the element position information bound to the second key frame may be updated based on the alignment transformation relationship.

[0096] In some embodiments, the element position information bound to the first key frame may be updated based on the alignment transformation relationship, and the element position information bound to the second key frame may be updated.

[0097] In some embodiments, based on the alignment transformation relationship, the element position information bound to the target key frame is updated, including: adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory; and updating the element position information bound to the target key frame according to the adjusted pose of the target key frame.

[0098] When adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory, a global pose optimization is required. When performing the global pose optimization, the pose of the target key frame needs to be adjusted based on the target constraint conditions.

[0099] The above-mentioned adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory may include but is not limited to: adjusting the pose of the target key frame based on the alignment transformation relationship and the target constraint condition to align the historical trajectory with the current trajectory;

[0100] The target constraint condition at least includes: a constraint condition on the relative posture between the first key frame and the second key frame.

[0101] In the present application, the constraint conditions of relative posture can be provided through the above alignment transformation relationship.

[0102] The above-mentioned target constraint conditions also include: a constraint condition of the recursive posture between the first key frame and adjacent key frames; and / or an error constraint condition between the position information of the first key frame and the corresponding positioning information of the first key frame.

[0103] For example, Figure 2A is a posture diagram. Figure 2A There are circles and lines in it. Figure 2A The circles in the figure represent vertices, and the lines represent edges. Assume that the key frames in the history trajectory include: , assuming that the key frames in the current trajectory include: , their edge constraints include timing edge constraints and loop edge constraints. The pose of each key frame in can be expressed as , for keyframes The pose of each key frame in can be expressed as ,in, =1, 2…8, =1, 2…4. Since there may be an erroneous matching relationship in the loop matching result, a loss function can be added when constructing the optimization problem. The loss function in this application can be of any type. The residual contribution and error corresponding to the following loss function are only examples.

[0104] The above alignment transformation relationship can be expressed as: , based on the alignment transformation relationship, the relative pose constraint can be provided, and the contribution of the loop edge to the residual can be obtained as follows (1):

[0105] (1)

[0106] In some embodiments, by adjusting and / or The target constraints that are satisfied may include:

[0107] is less than a first preset value, or, Take the minimum value among multiple adjustments.

[0108] Based on the constraints of the recursive pose between the first keyframe and the adjacent keyframes, the contribution of the temporal edge to the residual can be obtained as follows:

[0109] (2)

[0110] The absolute position based on the Global Positioning System (GPS) provides a one-dimensional edge constraint, and the error generated is the following formula (3):

[0111] (3)

[0112] in, Represents the absolute position of the Global Positioning System (GPS).

[0113] In some embodiments, by adjusting and / or The target constraints that are satisfied may include:

[0114] is less than the first preset value, or Take the minimum value among multiple adjustments.

[0115] In some embodiments, by adjusting and / or The target constraint condition that is satisfied may include at least one of the following:

[0116] (1) is less than a first preset value, or, Take the minimum value among multiple adjustments;

[0117] (2) is less than a second preset value, or, Take the minimum value among multiple adjustments;

[0118] (3) is less than a third preset value, or, Take the minimum value among multiple adjustments.

[0119] Based on the above method, the pose of the target keyframe can be adjusted based on the alignment transformation relationship to align the historical trajectory with the current trajectory. Afterwards, since each keyframe is bound to the feature position information, the feature position information bound to the target keyframe can be updated according to the adjusted pose of the target keyframe, thereby realizing the update of the features in the map.

[0120] Figure 2B This is a detailed flowchart of a feature update process. Figure 2B As shown, the element update process may include but is not limited to the following steps:

[0121] S111, determining whether it is a newly recorded route.

[0122] In other words, determine whether the current trajectory is a new route or an old route recorded in history.

[0123] In the present application, if the current track is a newly recorded route, the following step S112 may be directly executed based on the relevant information of the current track; if the current track is not a newly recorded route, the following step S114 may be executed.

[0124] S112, key frame selection and element binding.

[0125] For example, key frames can be selected from the trajectory at a certain distance, and the element position information corresponding to the trajectory within the certain distance is bound to the selected key frame. For example, assuming that the key frame pose matrix is , where R represents the vehicle direction, t represents the vehicle position; element shape point The representation after switching to the center of the keyframe is: .

[0126] S113. Data is stored in the database.

[0127] The positions and postures of the key frames, as well as the position information of the elements bound to the key frames and other data are stored in the database.

[0128] In some embodiments, the relationship between the update data (such as element location information) and the memory route may be maintained separately for each memory route (ie, each track).

[0129] In some embodiments, after selecting key frames for the trajectory and saving the key frame data, the road network can be segmented to divide the trajectory into intersection nodes (Node) and connecting sections (Skeleton) between intersections. When the data is updated, it can be updated in units of Node or Skeleton. At this time, it is only necessary to maintain the relationship table between the road network and the memory route.

[0130] S114: Determine the overlapping area of ​​geographical locations.

[0131] If the user drives along a previously remembered route again, the user can preliminarily screen the areas in the historical trajectory and the current trajectory that may have a loop relationship based on the geographic location. If the current trajectory only partially overlaps with the historical trajectory, only the overlapping area will be updated later.

[0132] S115: Determine whether the current route coincides with the learned route.

[0133] When determining whether the current route coincides with the learned route, it may be determined whether a first key frame in the historical trajectory and a second key frame in the current trajectory have a loop relationship.

[0134] When there is a loop relationship between the first key frame in the historical trajectory and the second key frame in the current trajectory, it is determined that the current route coincides with the learned route, and step S112 can be continued at this time; when there is no loop relationship between the first key frame in the historical trajectory and the second key frame in the current trajectory, it is determined that the current route does not coincide with the learned route, and step S114 can be returned to execute.

[0135] S116: Select candidate loop pairs.

[0136] If the position difference between the first key frame in the historical trajectory and the second key frame in the current trajectory is less than a preset value, it is determined that there is a loop relationship, and the first key frame and the second key frame can be determined as a candidate loop pair.

[0137] S117: Loopback matching.

[0138] Furthermore, point cloud alignment may be performed on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship.

[0139] S118. Global pose optimization.

[0140] In the global pose optimization process, the pose of the target keyframe is first adjusted based on the alignment transformation relationship to align the historical trajectory with the current trajectory; then, according to the adjusted pose of the target keyframe, the position information of the elements bound to the target keyframe is updated.

[0141] S119: The posture data is updated and re-stored.

[0142] The adjusted position and posture of the target key frame, as well as data such as the element position information bound to the updated target key frame can be stored in the database.

[0143] In the above-mentioned feature updating method, when there is a loop relationship between the first key frame in the current trajectory of the vehicle and the second key frame in the historical trajectory, it can be determined that the same route has been passed. At this time, according to the point cloud data corresponding to the recorded first key frame and the second key frame, the feature position information bound to the first key frame and / or the second key frame can be updated. In this way, when passing the same route again, the elements in the map can be updated, so that accurate feature information can be reconstructed during multiple driving processes, and the amount of data in each calculation process is small, which can be achieved through the vehicle side without uploading and sending data, reducing traffic overhead and algorithm complexity.

[0144] In the present application, before updating the element location information, an element reconstruction process may be performed first.

[0145] In an exemplary embodiment, Figure 3AFIG2 is a flow chart of an element updating method, which involves an element reconstruction process and an element updating process in a mapping process. The method includes the following steps S301 to S306:

[0146] S301. Acquire perception information collected by the vehicle on the target track.

[0147] The target trajectory includes: a historical trajectory or a current trajectory.

[0148] S302: Track and match elements based on the perception information to determine multiple associated position information of the elements.

[0149] For example, the reconstruction element categories may include, but are not limited to, lane lines, ground signs, and traffic sign poles. The ground signs may include, but are not limited to, arrows, stop lines, zebra crossings, and the like.

[0150] S303: Reconstruct the element position information bound to the target key frame based on the multiple associated position information of the element.

[0151] For example, take lane line reconstruction as an example. Figure 3B A schematic diagram of a lane reconstruction process is shown in FIG. Figure 3B As shown, the process may include the following steps:

[0152] S311, determine whether the KM matching is successful.

[0153] In some embodiments, based on the Kuhn-Munkres Algorithm (KM), target tracking and matching are performed based on the perception information to obtain associated data.

[0154] Exemplarily, taking the feature category of lane line as an example, during the feature tracking and matching process, it can be determined whether the lane lines belong to the same lane line based on the distance between the lane lines and the consistency of the attributes, that is, to determine multiple associated position information of the lane lines.

[0155] For example, Figure 4 A schematic diagram of the distance calculation between lane lines. Figure 4 As shown, Figure 4 Points A, B and C belong to the same lane line, and points D, E and F belong to the same lane line. During the feature tracking and matching process, the distance d1 between point A and point D, the distance d2 between point B and point E, and the distance d3 between point C and point F can be determined, and then it is determined whether the distance d1, the distance d2 and the distance d3 are less than the preset distance. If the distance d1, the distance d2 and the distance d3 are less than the preset distance, then the lane line to which points A, B and C belong, and the lane line to which points D, E and F belong can be the same lane line.

[0156] S312: Create a new tracking target.

[0157] S313, target association.

[0158] In this application, if the KM match is successful, it means that the current tracking target belongs to the same lane line as the existing lane line that has been matched and associated, that is, target association is performed. If the KM match is unsuccessful, it means that the current tracking target does not belong to the same lane line as the existing lane line that has been matched and associated. At this time, a new tracking target is created, that is, the new lane line is confirmed.

[0159] S314. The target disappears.

[0160] When the target leaves the field of view of the vehicle, subsequent steps S605 to S607 can be executed to restore the spatial 3D position, that is, element reconstruction.

[0161] S315, DBSCAN algorithm is used for aggregation.

[0162] S316: Calculate weights and cluster centers.

[0163] In some embodiments, reconstructing the element position information bound to the target key frame based on multiple associated position information of the element may include: reconstructing the element position information corresponding to the multiple associated position information based on the multiple associated position information of the element, and determining the element position information bound to the target key frame therefrom.

[0164] In some embodiments, based on multiple associated position information of an element, reconstructing the element position information bound to a target keyframe may include but is not limited to: determining multiple position information associated with each shape point in the element from multiple associated position information of the element; determining the position information of each shape point of the element based on multiple position information associated with each shape point in the element to reconstruct the element position information.

[0165] In some embodiments, multiple pieces of position information associated with each shape point in the element may be aggregated (clustered) to determine multiple pieces of position information associated with each shape point in the element.

[0166] In some embodiments, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm can be used. The DBSCAN algorithm is a density-based spatial clustering algorithm that is robust to noise. Multiple associated position information of elements are aggregated to achieve the above-mentioned point clustering and determine multiple position information associated with each point in the element.

[0167] For example, Figure 5This is a schematic diagram of a point aggregation. Figure 5 As shown, taking the feature category of lane line as an example, the points in each dotted ellipse box are multiple position information associated with each shape point of the aggregated lane line, and each dotted ellipse box is multiple observation points of the lane line belonging to the same shape point after clustering, and each point in each dotted ellipse box represents an observation point of the same lane line at different times.

[0168] In some embodiments, the position information of each shape point of the element is determined based on multiple position information associated with each shape point in the element to reconstruct the element position information. The position information of each shape point (that is, the above-mentioned cluster center) can be calculated by taking a weighted sum of the multiple position information associated with each shape point.

[0169] In some embodiments, based on multiple pieces of position information associated with each shape point in the element, determining the position information of each shape point of the element to reconstruct the element position information may include but is not limited to:

[0170] Determine the weight of multiple position information corresponding to each shape point; perform weighted summation on the multiple position information corresponding to each shape point in the element according to the weight of the multiple position information, determine the position information of each shape point of the element, and reconstruct the element position information;

[0171] The weight of each piece of location information in the plurality of location information is determined according to at least one of the following information:

[0172] The lateral distance between each position information and the vehicle;

[0173] The longitudinal distance between each position information and the vehicle;

[0174] The observation angle of the vehicle corresponding to each position information.

[0175] Among them, the weight of each position information is inversely correlated with the size of the horizontal distance; the weight of each position information is inversely correlated with the size of the longitudinal distance; the weight of each position information is inversely correlated with the size of the observation angle.

[0176] For example, the weight of each point in the cluster can be calculated according to the position and angle of each lane line from the vehicle when it is observed (e.g. Figure 4 The weight of each point in each dotted ellipse box in the above weight calculation formula can be the following formula (4):

[0177] (4)

[0178] in, is the longitudinal distance between the observation point (i.e. each position information corresponding to each of the above-mentioned shape points) and the vehicle, is the lateral distance between the observation point and the vehicle, yaw is the observation angle, a, b, c are the means of Gaussian distribution, and is the variance.

[0179] By using the weighted summation method, the position information of each shape point of the element is determined to reconstruct the element position information, which can effectively improve the position accuracy of the element (lane line) and remove the influence of external points.

[0180] S317, Bessel fitting.

[0181] In some embodiments, based on multiple position information associated with each shape point in the element, the position information of each shape point of the element is determined to reconstruct the element position information, which may include but is not limited to: based on multiple position information associated with each shape point in the element, the position information of each shape point of the element is determined; and the position information of multiple shape points of the element is fitted to reconstruct the element position information.

[0182] The fitting method may adopt a Bessel fitting method or any other fitting method to achieve a smoothing effect of the element (which may be a lane line).

[0183] S304: Determine whether a first key frame in the historical trajectory and a second key frame in the current trajectory have a loop relationship.

[0184] S305 , performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship.

[0185] S306: Based on the alignment transformation relationship, update the element position information bound to the target key frame.

[0186] For the description of the above steps S304 to S306 , reference may be made to the relevant description of the above steps S101 to S103 , which will not be repeated here.

[0187] In the above embodiment, the elements are first reconstructed for each driving trajectory. After the element reconstruction, when the same route is passed again, the elements in the map can be updated, so that accurate element information can be reconstructed during multiple driving processes. In addition, the amount of data in each calculation process is small, and incremental learning can be achieved through the vehicle side without uploading and sending data, thereby reducing traffic overhead and algorithm complexity.

[0188] In an exemplary embodiment, Figure 6 FIG3 is a flow chart of an element updating method, which involves an element updating process, and an element matching and fusion process in a mapping process. The method may include but is not limited to the following steps S601 to S606:

[0189] S601: Determine whether a first key frame in a historical trajectory and a second key frame in a current trajectory have a loop relationship.

[0190] S602: Perform point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship.

[0191] S603: Adjust the position and posture of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory.

[0192] S604: Update the element position information bound to the target key frame according to the adjusted position and posture of the target key frame.

[0193] For the description of the above steps S601 to S604 , reference may be made to the above related description of steps S101 to S103 , which will not be repeated here.

[0194] After updating the element position information through the above steps S601 to S604, element matching and fusion may be further performed based on the following steps S605 and S606.

[0195] S605: Acquire the aligned historical trajectory and the aligned current trajectory.

[0196] After the above step S603, the aligned historical trajectory and the aligned current trajectory may be saved, and at this time, the saved aligned historical trajectory and the aligned current trajectory may be obtained.

[0197] S606: Merge the element position information bound to the aligned historical track with the element position information bound to the aligned current track to obtain the element position information bound to the target track.

[0198] In some embodiments, the element position information bound to the aligned historical trajectory is fused with the element position information bound to the aligned current trajectory. The aligned current trajectory can be used as the basic trajectory, and the element position information bound to the aligned current trajectory and the element position information bound to the aligned historical trajectory are unified onto the basic trajectory. Then, element matching is performed, and the element position information is fused to finally obtain the element position information bound to the target trajectory.

[0199] In some embodiments, the element position information bound to the aligned historical trajectory is fused with the element position information bound to the aligned current trajectory. The aligned historical trajectory can be used as the basic trajectory, and the element position information bound to the aligned current trajectory and the element position information bound to the aligned historical trajectory are unified onto the basic trajectory. Then, element matching is performed, and the element position information is fused to finally obtain the element position information bound to the target trajectory.

[0200] After matching and fusing features and obtaining the feature location information bound to the target trajectory, topology post-processing and layer quality inspection can also be performed.

[0201] The element updating method in the embodiment of the present application is applied to an incremental self-learning mapping system. Figure 7 This is a flow chart of an incremental self-learning mapping system. Figure 7 As shown in FIG. 1 , the incremental self-learning mapping system includes the following Figure 7 Steps shown:

[0202] S701: Element reconstruction.

[0203] Among them, when reconstructing elements, the input is the relative position information of the perceived ground signs, lane lines, traffic signs, etc., as well as the absolute position information.

[0204] After reconstructing the elements based on the above input, the reconstructed element position information can be obtained.

[0205] For example, the reconstruction element categories may include, but are not limited to, lane lines, ground signs, and traffic sign poles. The ground signs may include, but are not limited to, arrows, stop lines, zebra crossings, and the like.

[0206] For the description of the above step S701, please refer to Figure 3A The relevant descriptions of steps S301 to S303 are not repeated here.

[0207] S702: Update elements.

[0208] In the present application, the reconstructed elements may be updated based on the element updating method provided in the embodiment of the present application.

[0209] For example, for the description of step S702, please refer to the above Figure 1 as well as Figure 2B The relevant description of the corresponding embodiment content will not be repeated here.

[0210] S703, element matching and fusion.

[0211] For example, for the description of step S702, please refer to the above Figure 6 The relevant descriptions of step S605 and step S606 are not repeated here.

[0212] S704: Topology post-processing.

[0213] In some embodiments, the above topology post-processing process can generate a road topology structure based on the reconstructed elements. For example, the road centerline is generated from the lane line according to the specification, and the topology binding logic generates a topological relationship for the generated road centerline, including the front, back, left, and right succession relationship of the centerline; the intersection range can be defined according to the zebra crossing and the stop line, and the intersection layer relationship can be generated; the relationship between the traffic light and the lane centerline can also be generated.

[0214] S705, layer quality inspection.

[0215] In some embodiments, the layer quality inspection process can evaluate the quality of the automatically generated map, check the topological self-consistency of each layer, the smoothness of the lane centerline, whether the memorized route can reach the destination, and set different severity levels according to different quality inspection items, and remind users in advance of problems that affect driving when using the map to ensure safety. The layer quality inspection process can also provide evaluation indicators for map completeness and display them on the map, so that users can intuitively feel that as the number of times the memorized route is learned increases, the resulting map becomes more accurate and detailed.

[0216] In this application, the mapping process involves steps such as feature reconstruction, feature update, feature matching and fusion, topology post-processing, and layer quality inspection. During the feature update process, the alignment of features is converted into trajectory alignment through key frame binding, which simplifies the problem difficulty, improves processing efficiency, and realizes incremental self-learning mapping on the vehicle side. The goal of improving the accuracy and completeness of the map as the number of route learning increases is achieved, giving users the feeling that the vehicle's autonomous driving on this route is becoming more and more intelligent.

[0217] Furthermore, the generated maps can be automatically inspected and evaluated for quality without the need for human intervention. Users can be alerted to problem areas in advance, allowing the autonomous driving system to automatically downgrade to problem areas, ensuring the safety of users' use of maps.

[0218] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0219] Based on the same inventive concept, the embodiment of the present application also provides a factor updating device for implementing the above-mentioned factor updating method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in the one or more factor updating device embodiments provided below can refer to the limitations on the factor updating method above, and will not be repeated here.

[0220] In an exemplary embodiment, Figure 8 The structural block diagram of an element updating device includes:

[0221] A determination module 801 is used to obtain a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory;

[0222] Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship;

[0223] The updating module 802 is used to update the element position information bound to the target key frame based on the alignment transformation relationship, and the target key frame includes: the first key frame and / or the second key frame.

[0224] In one embodiment, the updating module 802 is specifically used to update the element position information bound to the target key frame based on the alignment transformation relationship, including:

[0225] The pose of the target key frame is adjusted based on the alignment transformation relationship to align the historical trajectory with the current trajectory; and the element position information bound to the target key frame is updated according to the adjusted pose of the target key frame.

[0226] In one embodiment, the updating module 802 is specifically used to adjust the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory, including:

[0227] Based on the alignment transformation relationship and the target constraint condition, adjusting the pose of the target key frame to align the historical trajectory with the current trajectory;

[0228] The target constraint condition includes at least: a constraint condition on the relative posture between the first key frame and the second key frame.

[0229] In one embodiment, the target constraint condition further includes:

[0230] Constraints on the recursive pose between the first key frame and adjacent key frames;

[0231] and / or,

[0232] An error constraint condition between the position information of the first key frame and the corresponding positioning information of the first key frame.

[0233] In one embodiment, the element updating device further includes:

[0234] The fusion module 803 is used to obtain the aligned historical trajectory and the aligned current trajectory after the updating module 802 updates the element position information bound to the target key frame according to the adjusted position of the target key frame; and fuse the element position information bound to the aligned historical trajectory with the element position information bound to the aligned current trajectory to obtain the element position information bound to the target trajectory.

[0235] In one embodiment, the element updating device further includes:

[0236] The reconstruction module 804 is used to determine that the module 801 is also used to obtain the perception information collected by the vehicle on the target trajectory, and the target trajectory includes: the historical trajectory, or the current trajectory; perform element tracking and matching based on the perception information to determine multiple associated position information of the elements; and reconstruct the element position information bound to the target key frame based on the multiple associated position information of the elements.

[0237] In one embodiment, the reconstruction module 804 is specifically configured to:

[0238] The method of reconstructing the element position information bound to the first key frame and the second key frame based on the multiple associated position information of the element includes: determining the multiple position information associated with each shape point in the element from the multiple associated position information of the element; determining the position information of each shape point of the element based on the multiple position information associated with each shape point in the element, so as to reconstruct the element position information.

[0239] In one embodiment, the reconstruction module 804 is specifically used to: determine the position information of each shape point of the element according to the multiple position information associated with each shape point in the element to reconstruct the element position information, including: determining the weight of the multiple position information corresponding to each shape point; performing weighted summation on the multiple position information corresponding to each shape point in the element according to the weight of the multiple position information to determine the position information of each shape point of the element to reconstruct the element position information;

[0240] The weight of each piece of location information in the plurality of pieces of location information is determined according to at least one of the following information:

[0241] a lateral distance between each piece of position information and the vehicle;

[0242] a longitudinal distance between each piece of position information and the vehicle;

[0243] The observation angle of the vehicle corresponding to each piece of position information.

[0244] In one embodiment, the reconstruction module 804 is specifically used to: determine the position information of each shape point of the element according to multiple position information associated with each shape point in the element to reconstruct the element position information, including: determine the position information of each shape point of the element according to multiple position information associated with each shape point in the element; fit the position information of multiple shape points of the element to reconstruct the element position information.

[0245] Each module in the above-mentioned element updating device can be implemented in whole or in part by software, hardware and a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module above.

[0246] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, for example, a vehicle-mounted terminal, and its internal structure diagram may be as shown in FIG. Fig. 9As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for updating an element is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0247] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0248] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, each process of the factor updating method shown in the above embodiment is implemented.

[0249] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the factor updating method shown in the above embodiment are implemented.

[0250] In one embodiment, a computer program product is provided, including a computer program, which implements the various processes of the factor updating method shown in the above embodiment when executed by a processor.

[0251] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0252] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0253] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for updating an element, characterized in that: The method comprises: Obtaining a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory; Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship; Based on the alignment transformation relationship, updating the element position information bound to the target key frame, the target key frame including: the first key frame, and / or the second key frame; The updating of the element position information bound to the target key frame based on the alignment transformation relationship includes: Adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory; According to the adjusted position and posture of the target key frame, the element position information bound to the target key frame is updated.

2. The method according to claim 1, characterized in that The adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory includes: Based on the alignment transformation relationship and the target constraint condition, adjusting the pose of the target key frame to align the historical trajectory with the current trajectory; The target constraint condition includes at least: a constraint condition on the relative posture between the first key frame and the second key frame.

3. The method according to claim 2, characterized in that The target constraints also include: Constraints on the recursive pose between the first key frame and adjacent key frames; and / or, An error constraint condition between the position information of the first key frame and the corresponding positioning information of the first key frame.

4. The method according to claim 1, characterized in that: After updating the element position information bound to the target key frame according to the adjusted position of the target key frame, the method further includes: Acquire the aligned historical trajectory and the aligned current trajectory; The element position information bound to the aligned historical track is merged with the element position information bound to the aligned current track to obtain the element position information bound to the target track.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Acquire perception information collected by the vehicle on a target trajectory, wherein the target trajectory includes: the historical trajectory, or the current trajectory; Perform element tracking and matching based on the perception information to determine multiple associated position information of the elements; Based on the multiple associated position information of the elements, the element position information bound to the target key frame is reconstructed.

6. The method according to claim 5, characterized in that The step of reconstructing the element position information bound to the target key frame based on the plurality of associated position information of the element includes: Determine, from the plurality of associated position information of the element, a plurality of position information associated with each shape point in the element; According to a plurality of position information associated with each shape point in the element, the position information of each shape point of the element is determined to reconstruct the element position information.

7. The method according to claim 6, characterized in that Determining the position information of each shape point of the element according to the plurality of position information associated with each shape point in the element to reconstruct the element position information includes: Determine the weight of the plurality of position information corresponding to each shape point; For each shape point in the element, weighted sum is performed on the multiple position information corresponding to the multiple position information according to the weight of the multiple position information to determine the position information of each shape point of the element, so as to reconstruct the element position information; The weight of each piece of location information in the plurality of pieces of location information is determined according to at least one of the following information: a lateral distance between each piece of position information and the vehicle; a longitudinal distance between each piece of position information and the vehicle; The observation angle of the vehicle corresponding to each piece of position information.

8. The method according to claim 5, characterized in that Determining the position information of each shape point of the element according to the plurality of position information associated with each shape point in the element to reconstruct the element position information includes: Determine the position information of each shape point of the element according to the plurality of position information associated with each shape point in the element; The position information of the plurality of shape points of the element is fitted to reconstruct the position information of the element.

9. A factor updating device, characterized in that: The device comprises: A determination module, configured to obtain a first key frame in a historical trajectory and a second key frame in a current trajectory, wherein the first key frame and the second key frame are bound with element position information corresponding to the trajectory; Performing loop detection on the first key frame and the second key frame, and when detecting that the first key frame and the second key frame have a loop relationship, performing point cloud alignment on the point cloud data corresponding to the first key frame and the point cloud data corresponding to the second key frame to determine an alignment transformation relationship; An updating module, configured to update the element position information bound to a target key frame based on the alignment transformation relationship, wherein the target key frame includes: the first key frame and / or the second key frame; The updating of the element position information bound to the target key frame based on the alignment transformation relationship includes: Adjusting the pose of the target key frame based on the alignment transformation relationship to align the historical trajectory with the current trajectory; According to the adjusted position and posture of the target key frame, the element position information bound to the target key frame is updated.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images