Change discovery method, device and storage medium for ground features

Abstract

Embodiments of the present specification provide a method and device for discovering changes in geographic features and a storage medium. The method comprises: obtaining new acquisition data of a target geographic area, the new acquisition data comprising new trajectory data and new image data; associating the new trajectory data with the new image data to obtain new trajectory-image association data; obtaining historical trajectory-image association data of the target geographic area; and determining whether there is a changed geographic feature in the target geographic area based on the new trajectory-image association data and the historical trajectory-image association data.

Description

Technical Field

[0001] This specification relates to the field of map technology, and in particular to methods, devices and storage media for detecting changes in geographic features. Background Technology

[0002] Currently, vehicle-mounted driver assistance systems (ADAS), advanced driver assistance systems (ADAS), and smart city construction all rely on high-precision maps (HMRs). Building an HMR involves first collecting data from various sensors on the vehicle, including image data and point cloud data. Then, based on the collected data, various types of high-precision geographic feature data (such as lanes and signs) are generated, ultimately forming the HMR.

[0003] As time changes, real-world geographical features also change. To ensure that the geographical features in high-definition maps remain consistent with the real world, these maps need to be updated regularly. Current update methods typically involve a map-collecting vehicle gathering new data in the area requiring updating, and then creating a new high-definition map based on this new data. In other words, existing technology involves processing all new data to obtain a new high-definition map. However, as is well known, changes in the real world are generally localized; therefore, processing all new data is inefficient and costly. Thus, improving the efficiency and reducing the cost of high-definition map updates is a technical problem that those skilled in the art need to solve. Summary of the Invention

[0004] To overcome the problems existing in related technologies, this specification provides methods, devices and storage media for discovering changes in ground features.

[0005] According to a first aspect of the embodiments of this specification, a method for detecting changes in ground features is provided, the method comprising:

[0006] Acquire newly collected data for the target geographic area, including new trajectory data and new image data;

[0007] The new trajectory data is associated with the new image data to obtain new trajectory image associated data;

[0008] Obtain historical trajectory image association data of the target geographic area;

[0009] Based on the new trajectory image association data and the historical trajectory image association data, it is determined whether there are any changed geographical features in the target geographic area.

[0010] According to a second aspect of the embodiments of this specification, a device for detecting changes in ground features is provided, the device comprising:

[0011] The new data acquisition module is used to: acquire newly collected data of the target geographic area, the newly collected data including new trajectory data and new image data;

[0012] The association module is used to: associate the new trajectory data with the new image data to obtain new trajectory image association data;

[0013] The historical data acquisition module is used to: acquire historical trajectory image association data of the target geographic area;

[0014] The change determination module is used to: determine whether there are any changed geographical features in the target geographic area based on the new trajectory image association data and the historical trajectory image association data.

[0015] According to a third aspect of the embodiments of this specification, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method for discovering changes in ground features as described in any of the preceding aspects.

[0016] According to a fourth aspect of the embodiments of this specification, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method embodiment for discovering changes in ground features described in the first aspect.

[0017] The technical solutions provided in the embodiments of this specification may include the following beneficial effects:

[0018] In this embodiment, the newly acquired data for the target geographic area includes new trajectory data and new image data. New trajectory image association data is obtained by associating the new trajectory data and new image data. Similarly, historical trajectory image association data for the target geographic area is acquired. Based on this, a comparison between the new trajectory image association data and the historical trajectory image association data can determine whether there are any changed geographic features in the target geographic area. Because this embodiment can detect changes in geographic features, it does not require full processing of the newly acquired data, thereby reducing the update cost of high-precision maps and improving their update efficiency.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this specification. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this specification and, together with the description, serve to explain the principles of this specification.

[0021] Figure 1A This is a flowchart illustrating a method for detecting changes in geographic features according to an exemplary embodiment.

[0022] Figure 1B This is a schematic diagram of a road surface frame according to an exemplary embodiment of this specification.

[0023] Figure 2A This is a schematic diagram illustrating the processing of newly acquired data according to an exemplary embodiment of this specification.

[0024] Figure 2B This is a schematic diagram of a client according to an exemplary embodiment of this specification.

[0025] Figure 2C This is a schematic diagram illustrating the processing of another newly acquired data according to an exemplary embodiment of this specification.

[0026] Figure 3 This is a structural diagram of a computer device housing a change detection device for geographic features, as illustrated in this specification according to an exemplary embodiment.

[0027] Figure 4 This is a block diagram illustrating a change detection device for ground features according to an exemplary embodiment. Detailed Implementation

[0028] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this specification as detailed in the appended claims.

[0029] The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of this specification. The singular forms “a,” “the,” and “the” as used in this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0030] It should be understood that although the terms first, second, third, etc., may be used in this specification to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this specification, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0031] In some solutions, map data collection vehicles gather new data for scenarios requiring updates, and then create new high-precision maps based on this new data. For example, a 5000km highway was surveyed in June 2021. In January 2022, to update the changes to this highway over the past six months, a second data collection vehicle was deployed to survey the same highway. The second data collection involved full processing of the new data. However, statistics show that the real-world change rate is only around 5%. This means that only about 250km of the changed section actually required map data modification, while 4750km of data resources were wasted. Clearly, in update scenarios, this solution results in significant resource waste, relatively low efficiency, and excessively long delivery cycles due to redundant data processing and repetitive work. This problem will become increasingly apparent as the frequency of data collection and updates increases.

[0032] Based on this, this specification provides a method for detecting changes in geographic features. The newly acquired data for the target geographic area includes new trajectory data and new image data. New trajectory image association data is obtained by associating the new trajectory data and the new image data. Similarly, historical trajectory image association data for the target geographic area is acquired. Based on this, a comparison between the new trajectory image association data and the historical trajectory image association data can determine whether there are any changed geographic features in the target geographic area. Since this embodiment can detect changes in geographic features, it does not require full processing of the newly acquired data, thereby reducing the update cost of high-precision maps and improving the update efficiency. The embodiments of this specification will now be described in detail.

[0033] like Figure 1A As shown, Figure 1A This is a flowchart illustrating a method for detecting changes in ground features according to an exemplary embodiment, comprising the following steps:

[0034] In step 102, new data collection is obtained for the target geographic area.

[0035] The newly acquired data includes new trajectory data and new image data.

[0036] In step 104, the new trajectory data is associated with the new image data to obtain new trajectory image associated data.

[0037] In step 106, historical trajectory image association data of the target geographic area is obtained.

[0038] In step 108, based on the new trajectory image association data and the historical trajectory image association data, it is determined whether there are any changed geographical features in the target geographic area.

[0039] The data acquisition device in this embodiment can be a mobile device, such as a vehicle. As an example, a map acquisition vehicle is a professional data acquisition device used to create high-precision maps, and it is usually equipped with data acquisition devices such as LiDAR, image acquisition devices, positioning modules, and inertial measurement units (IMUs).

[0040] For example, while the map data collection vehicle is moving, the positioning module can collect the vehicle's geographical location information, namely longitude and latitude, based on the Global Navigation Satellite System (GNSS). The positioning module collects geographical location information according to a first predetermined cycle, obtaining a trajectory point each time. The trajectory points collected multiple times form trajectory data, which carries geographical location information and time information. In this embodiment, the new trajectory data is a sequence of trajectory points composed of multiple trajectory points.

[0041] Similarly, as the map acquisition vehicle moves, the image acquisition device also acquires images according to the second preset cycle. The images acquired multiple times form image data. Each frame of the image can carry time information, and optionally, it can also carry geographical location information acquired by the positioning module built into the image acquisition device, i.e., the location where the image was captured, and attitude information acquired by the IMU module built into the image acquisition device, etc. This embodiment does not limit this. In this embodiment, the positioning module and the image acquisition device of the map acquisition vehicle work independently, that is, their preset acquisition cycles are different.

[0042] Similarly, other devices on the map collection vehicle also work independently. For example, the lidar independently collects point clouds to form point cloud data, which also carries time information, attitude information, or geographical location information, etc.

[0043] The trajectory data collected by the positioning module, the image data collected by the image acquisition device, and / or the point cloud data collected by the lidar are collectively referred to as data data. It is understood that in practical applications, depending on the configuration of the map acquisition vehicle and the actual acquisition needs, the data data may also include other types of data, and this embodiment does not limit this.

[0044] By collecting data, real-world features such as various road elements collected by the map-collecting vehicle can be identified. These features can include lane lines, stop lines, yield lines, ground arrows, text, guide lines, directional poles, signs, traffic lights, traffic signs, gantries, buildings, etc. In practical applications, these features can be configured according to the needs of high-precision map production; this embodiment does not impose any limitations on this. For example, after identifying features from the data, the characteristic information of each feature can be extracted to construct a high-precision map. This can be achieved by extracting image features from image data, extracting point cloud features from point cloud data, or fusing image data and point cloud data to extract features.

[0045] After a high-precision map of a certain geographical area is created, the data used to create the map is stored as historical data in the database. For ease of distinction, this embodiment refers to the historically collected data used to create high-precision maps as historical collected data, and the data collected for a certain geographical area to be used to create a high-precision map as newly collected data for the target geographical area. Those skilled in the art will understand that in practical applications, the scope of the target geographical area can be flexibly configured as needed, and this embodiment does not limit this.

[0046] As mentioned earlier, the newly acquired data includes various types of data. In this embodiment, to identify whether ground features have changed, the identification is based on new trajectory data and new image data from the newly acquired data. Compared to point cloud data, image data has a relatively smaller data volume, and it can more quickly extract accurate and rich visual features of ground features in the real world. Furthermore, this embodiment associates the new trajectory data with the new image data, and combined with subsequent processing, it can accurately identify ground features that have changed in the target geographic area.

[0047] The association between the new trajectory data and the new image data can be based on the time and / or geographical location information of the trajectory points in the new trajectory data, and the time and / or shooting location of the images in the new image data. For example, a new image with matching locations can be associated with a new trajectory point, or a new image with matching times can be associated with a new trajectory point, or a new image with matching locations and times can be associated with a new trajectory point, and so on. Taking location matching as an example, there are multiple ways to match the locations of the new image and the new trajectory point. For instance, the shooting location of the new image and the geographical location of the new trajectory point can be completely consistent, or the difference between the two can meet a matching condition. For example, a match is determined when the difference between the two is less than a preset difference threshold, and the preset difference threshold can be configured according to actual needs.

[0048] The association can be implemented by recording the association relationship between the identifier of the new image and the identifier of the new trajectory point, or by configuring the new image and the new trajectory point with the same marker, etc. Those skilled in the art will know that it can be implemented as needed in actual applications, and this embodiment does not limit it.

[0049] In some cases, due to differences in the acquisition period of trajectory points and the acquisition period of images, some new images in the new image data may have trajectory points in the new trajectory data that perfectly match their locations. However, there may also be cases where the shooting locations of some new images in the new image data do not have trajectory points in the new trajectory data that perfectly match their shooting locations. As in the previously described embodiment, there may be some differences between the shooting locations of the new images and the geographical locations of the new trajectory points. If the differences are small, the two can be correlated. In this case, there will be a certain degree of error in the correlation data of the new trajectory images.

[0050] To obtain new trajectory image data with higher accuracy, in this embodiment, the new trajectory data is associated with the new image data to obtain new trajectory image associated data, which may include:

[0051] New image trajectory points are added to the new trajectory data, and the new images in the new image data are associated with the new image trajectory points to obtain new trajectory image association data.

[0052] For example, trajectory points can be added to the new trajectory data based on the geographic location information of the new image, so that for each new image in the new image data, there is a new image trajectory point in the new trajectory data that matches its location. Based on this, for each new image in the new image data, there is a trajectory point associated with it in the new trajectory data, thereby obtaining highly accurate new trajectory image association data. When performing subsequent change identification, the accuracy of detecting changes in ground features can also be guaranteed.

[0053] In some cases, the acquisition cycle of trajectory points is short, that is, the number of trajectory points acquired is very large and the density is high. In order to improve data processing efficiency, the method of this embodiment may further include: thinning the trajectory points in the new trajectory data according to a set rule to obtain thinned new trajectory data.

[0054] The step of adding new image trajectory points to the new trajectory data and associating the new images in the new image data with the new image trajectory points to obtain new trajectory image association data may include: adding new image trajectory points to the thinned new trajectory data and associating the new images in the new image data with the new image trajectory points to obtain new trajectory image association data.

[0055] The thinning rules can be implemented in various ways. For example, they can include a rule of M trajectory points every N meters. N and M can be set arbitrarily. Those skilled in the art know that specific thinning rules can be flexibly implemented in practical applications. This embodiment is not limited in this respect.

[0056] In practical applications, the data can be thinned first, and then new image trajectory points can be added to the thinned trajectory data. The new images in the new image data are then associated with these new image trajectory points to obtain new trajectory image-associated data. This method improves data processing efficiency by thinning the high-density new trajectory data, and then adding image trajectory points ensures that every new image in the new image data has an associated trajectory point in the new trajectory data.

[0057] In practical applications, there are several ways to add new image trajectory points. In some examples, adding image trajectory points to the thinned new trajectory data and associating the new image in the new image data with the new image trajectory points to obtain new trajectory image association data can include:

[0058] The shooting location of the new image, including the new image data, is added to the thinned new trajectory data as the new image trajectory point corresponding to the new image;

[0059] The new image is associated with its corresponding new image trajectory points to obtain new trajectory image association data.

[0060] In this embodiment, new image trajectory points can be added to the thinned new trajectory data by utilizing the shooting location of the new image. Then, the new image is associated with its corresponding new image trajectory points to obtain new trajectory image association data. In this way, the positions of the new image and the new image trajectory points are perfectly matched, and the accuracy of the obtained new trajectory image association data is very high, which can ensure that the subsequent accurate determination of whether the ground features have changed is guaranteed.

[0061] In other examples, adding new image trajectory points to the thinned new trajectory data, and associating the new images in the new image data with the new image trajectory points to obtain new trajectory image association data, may include:

[0062] Based on the new trajectory points included in the thinned new trajectory data, interpolation processing is performed to obtain the positions of the new image trajectory points;

[0063] The new image data includes matching the shooting location of the new image with the location of the trajectory point of the new image, and then associating the new image with the matching location with the trajectory point of the new image to obtain new trajectory image association data.

[0064] This embodiment performs interpolation processing on the thinned new trajectory data. The interpolation rules can be flexibly configured as needed. For example, the interpolation method can be determined according to the acquisition cycle of the new image, so that the new trajectory points in the interpolated new trajectory data match the new image as closely as possible. After obtaining the image trajectory points through interpolation processing, the shooting position of the new image included in the new image data can be matched with the position of the new image trajectory points. The new image with the matched position is associated with the new image trajectory points to obtain new trajectory image association data.

[0065] The matching of the new image's capture location with the location of the image trajectory points can be achieved in various ways. It can be that the two locations are identical, or that the difference between the two locations meets a matching condition. For example, a match is determined when the difference between the two locations is less than a preset difference threshold. Based on this, this embodiment can improve data processing efficiency and obtain highly accurate new trajectory image association data, ensuring accurate subsequent determination of whether ground features have changed.

[0066] Similarly, historical trajectory image association data can be obtained in the same way. The target geographic area includes historical data collected in the past, which has been used to create high-precision maps. This historical data also includes historical trajectory data, historical image data, and other types of historical data. By associating historical trajectory data and historical image data, historical trajectory image association data is obtained.

[0067] In some cases, considering that map data collection vehicles typically travel long distances along roads, the amount of newly collected data for the target geographic area is large, covering a wide geographical area. Therefore, in this embodiment, the method may further include: segmenting the newly collected data for the target geographic area to obtain segmented and stored new data.

[0068] For example, the segmentation process can be implemented in various ways. For instance, it can be a fixed-length segmentation rule, where each segment covers a distance of S meters, and S can be configured as needed, such as 200 meters, 300 meters, 500 meters, or 1000 meters. Alternatively, the geographical coverage of each segment can also be different, and can be configured according to actual business needs; this embodiment does not impose any limitations on this. Segmentation processing can improve data processing efficiency.

[0069] In other examples, the segmentation of newly collected data for the target geographic area to obtain segmented and stored newly collected data includes:

[0070] Obtain a pre-made road surface bounding box, which represents the three-dimensional spatial information of the road. The road surface bounding box is made based on road data from an existing high-precision map, which is related to the historical trajectory image association data.

[0071] The road surface bounding box is used to determine the road to which the new trajectory data in the newly collected data belongs;

[0072] The newly collected data for the target geographic area is segmented based on the road to which the new trajectory data belongs, resulting in segmented and stored new data.

[0073] In practical applications, some roads are located close to each other, such as main roads and auxiliary roads, with the auxiliary road being closer to the main road. Therefore, it is necessary to accurately determine the road where the map data collection vehicle is located when collecting trajectory data, so that new trajectory data can be accurately matched with historical trajectory data subsequently.

[0074] In practical applications, established high-precision maps include accurate road data, which can be used to describe the precise three-dimensional representation of each road. This includes road geometric structure information, data attribute information of each lane (geographical location, slope, curvature, heading, elevation, etc.), shape information of road markings, and information such as road medians. Based on this, road surface frames corresponding to each road can be pre-generated using the road data in the existing high-precision map. The road surface frames represent the three-dimensional spatial information of the corresponding road. For example, the road surface frame can be a solid model built based on the road surface, carrying the road's longitude, latitude, and elevation information. Its specific shape is determined according to the actual road geometry. The method of generating the road surface frames can refer to relevant technologies, and this embodiment does not limit it. Figure 1B The diagram shows two roads, Road A (represented by grayscale blocks in the diagram, which includes three lanes) and Road B, along with the corresponding road surface boxes for each road.

[0075] For example, each trajectory point in the newly acquired trajectory data carries three-dimensional spatial information. The road to which each trajectory point belongs can be determined based on the relative positional relationship between the three-dimensional spatial information of each trajectory point and the road frame. Similarly, the road to which each historical trajectory point belongs can also be determined based on the relative positional relationship between the three-dimensional spatial information of the historical trajectory point and the road frame. Therefore, by using the road frame to determine the road to which the new trajectory data in the newly acquired data belongs, and by segmenting the newly acquired data for the target geographic area according to the road to which the new trajectory data belongs, segmented newly acquired data can be obtained. Based on this, for a segment of trajectory data in the newly acquired trajectory data, matching historical trajectory data can be accurately found. For example, a segment of data corresponds to a section of an actual road, and subsequent change detection can be performed segment by segment, such as performing subsequent steps in parallel on multiple routes; furthermore, because the segmented processing divides the data into smaller granularities, it is also easier to determine which location in which segment of the data shows changes in ground features, facilitating subsequent small-granular updates to the established high-precision map, thereby improving map update efficiency.

[0076] Newly acquired data for the target geographic area is segmented to obtain segmented, stored new data. Optionally, segmented new trajectory image association data can be obtained. Similarly, historically acquired data is also segmented to obtain segmented, stored historical data, which can also yield segmented new trajectory image association data. Therefore, when determining whether there are changed geographic features in the target geographic area, the process can be performed segment by segment. For example, for a segment of newly acquired data, a corresponding segment of new trajectory image association data is obtained. A matching segment of historical trajectory image association data is then retrieved from the historically acquired data. The matching process can involve comparing the positions of trajectory points in the new trajectory data with the positions of historical trajectory points in the historical trajectory data.

[0077] In some examples, based on the new trajectory image association data and the historical trajectory image association data, it is determined whether there are changed geographic features in the target geographic area, including:

[0078] The new image trajectory points and historical image trajectory points with matching locations are obtained from the new trajectory image association data and the historical trajectory image association data. The new image associated with the new image trajectory points and the historical image associated with the historical image trajectory points are compared. Based on the comparison results, it is determined whether there are any ground features whose positions have changed.

[0079] For example, the positions of each new image trajectory point in the new trajectory image association data are matched with the positions of each historical image trajectory point in the historical trajectory image association data. For a pair of new image trajectory points and historical image trajectory points that match the positions, the new image associated with the new image trajectory point and the historical image associated with the historical image trajectory point are compared. Based on the comparison results, it is determined whether there are any ground features whose positions of the new image trajectory point have changed.

[0080] In some cases, image matching algorithms can be used to calculate whether there are differences between new and historical images, and to determine the content of the differences. Considering that in practical applications, the number of changing ground features in the real world may be small and infrequent, and image matching algorithms may struggle to achieve high accuracy, in other cases, to improve recognition accuracy, new and historical images can be sent to the client so that the client can view the new and historical images and obtain the user-inputted change status.

[0081] As an example, the newly acquired trajectory image associated data and the corresponding historical trajectory image associated data obtained by segmenting the data in the aforementioned embodiment can be used as a task, and the image can be manually reviewed to see if there are any changes between the image and the historical image in the historical sub-data. In this embodiment, by segmenting the data, each segment of data can be sent to the client as a review task, and multiple users can review multiple segments of data at the same time, thereby improving data processing efficiency.

[0082] As an example, the client displays input objects for user input. For instance, the client displays input controls that the user can trigger to change. In response to the user triggering the input control, the client obtains the change status of the user-inputted feature representing the change. Optionally, the input control can be a control that allows the user to input changes, meaning the user only needs to trigger the input control when there is a change, thereby reducing the user's input operations.

[0083] Depending on whether the ground features have changed, different processing can be performed; for example, the method further includes: if it is determined that the location of the new image trajectory point has changed ground features, it is determined that the location of the new image trajectory point needs to be updated on the high-precision map; if it is determined that the location of the new image trajectory point has not changed ground features, the new image trajectory point and its associated new image are stored.

[0084] In this embodiment, for locations where ground features have changed, a high-precision map needs to be updated. Data collected at these locations can be used to create a local high-precision map, enabling a partial update of the existing high-precision map. For example, image trajectory points can be marked, indicating that the location of these points has changed ground features. Subsequently, all image trajectory points marked with this mark can be acquired before updating the high-precision map.

[0085] If the location of a new image trajectory point does not change, the new image trajectory point and its associated new image can be stored. Other data such as point cloud data corresponding to the location can also be stored in the database. There is no need to update the high-precision map for that location, thereby reducing processing costs and improving the update efficiency of the high-precision map.

[0086] In practical applications, there may be situations where some locations within the target geographic area have historically had no data collected, while the newly collected data from the map acquisition vehicle covers those locations. Therefore, in some examples, the method may further include: for the new image trajectory points included in the new trajectory image association data, searching in the historical trajectory image association data for historical image trajectory points with matching locations.

[0087] For example, the locations of image trajectory points included in the new trajectory image association data can be searched against historical trajectory image association data to see if there are any historical trajectory points matching those locations. If no matching points are found, it indicates that there is no historical data for that location, nor has a high-precision map been created there. Therefore, it can be determined that a high-precision map needs to be created for the location of the new image trajectory point. For example, the image trajectory point can be marked, indicating that a high-precision map needs to be created for that location. Subsequently, all image trajectory points with this mark can be acquired before creating the high-precision map. Therefore, high-precision maps can be created for localized areas without requiring full processing of the newly acquired data, reducing processing costs and improving the efficiency of high-precision map updates.

[0088] In practical applications, the creation of high-precision maps may take a considerable amount of time. For example, various processing steps are required before creating a high-precision map, such as projecting image data and point cloud data, generating and storing feature information of ground features from point clouds and / or image data, etc. The process of generating feature information usually takes a considerable amount of time. Based on this, in some other examples, the historical data collected includes historical point cloud data corresponding to the historical image data. The method further includes: acquiring historical feature data, which includes: historical feature information generated using historical image data and / or the historical point cloud data corresponding to historical image trajectory points in the historical trajectory image association data; and searching the historical feature data for any location-matching historical feature information for new image trajectory points included in the new trajectory image association data.

[0089] For example, the correspondence between historical point cloud data and historical image data can be determined by using the timestamps of historical image acquisition and point cloud acquisition. Since historical trajectory points in historical trajectory image association data are associated with historical images, and historical images are associated with historical point clouds, historical trajectory points correspond to historical image data and / or historical feature information generated from the historical point cloud data. Therefore, this embodiment can search the historical feature data for any matching historical feature information for new image trajectory points included in the new trajectory image association data. Different processing can be performed based on different search results. For example, if a historical trajectory point corresponding to a new image trajectory point exists, but no historical feature information is found, it can be determined that the generation of historical feature information at that location has not yet been completed and may be in the process of generation. The corresponding processing flow can be determined by whether the geographical features of the location have changed. For example, if there has been a change, a high-precision map can be built based on the historical feature information of the location and the newly collected data corresponding to the location after the historical feature information is generated; or the existing feature information generation process can be interrupted and a high-precision map corresponding to the geographical location can be built based on the newly collected data. If there has been no change, the newly collected data can be stored without affecting the existing process, etc.

[0090] The following description uses another embodiment. Figure 2A The diagram shown is a schematic representation of the processing of newly acquired data according to an exemplary embodiment of this specification. In this embodiment, after the data is acquired and transmitted back from the acquisition device, it can be processed as follows: Figure 2AThe solution outlines four steps: data standardization, identification of changes, data classification and aggregation, and data classification processing. This embodiment can identify and classify raw data, with different update processes corresponding to different changes. In particular, data showing no changes in identified geographic features will not undergo subsequent full-process processing and manual intervention, significantly improving the efficiency of high-precision map production and reducing production costs.

[0091] As an example, during data standardization, data can be standardized based on the data transmitted back from the data collection vehicle, providing standardized data input for subsequent processes. Standardization can refer to the data standardization process used in high-precision map production, including but not limited to reliability evaluation of trajectory points, trajectory segmentation, point cloud upload waiting, point cloud computation, image computation, image format conversion, thinning, or interpolation.

[0092] As an example, newly acquired data can be segmented. For instance, based on the location of trajectory points in the new trajectory data, multiple trajectory points can be divided to obtain fine-grained data. Point cloud processing, image processing, and format conversion can output standardized format point clouds, trajectory points, and images. Thinning can be performed before or after the new trajectory data is segmented. For example, according to a set rule, one or more trajectory points can be extracted from the new trajectory data every N meters. The specific N and the number of trajectory points extracted can be flexibly determined according to business needs. For example, a 500-meter segment could be used as a reference, and a trajectory point could be extracted every 100 meters within that segment.

[0093] In this embodiment, new trajectory data and new image data are correlated to obtain new trajectory image-correlated data. The correlation process can be referred to the previous embodiment. After correlation, each image trajectory point in the new trajectory image-correlated data corresponds to one or more images collected by the acquisition vehicle at that location. The number of images is determined based on the number of camera devices mounted on the acquisition vehicle. For example, the acquisition vehicle is equipped with four camera devices, corresponding to four different directions, such as front, back, left, and right, taking the vehicle's direction as an example. Thus, one trajectory point can correspond to four images. Each segment of newly acquired data includes multiple new image trajectory points in the thinned new trajectory data, as well as one or more images corresponding to each new image trajectory point.

[0094] After the above processing, differential identification can be performed. In this embodiment, the change status of each newly acquired data segment relative to the historical data segment can be determined based on each segment of newly acquired data and historical data matching its location. In some examples, there can be multiple types of change status. For example, some geographical areas do not have high-precision maps, and the newly acquired data covers geographical areas where high-precision maps are not established; therefore, the change status can include whether a high-precision map has been established. In other examples, the change status can also include whether the features of the ground features in the newly acquired data have changed. In other examples, the geographical area corresponding to the newly acquired data has a high-precision map, but the following situation may exist: feature information of ground features needs to be extracted before establishing a high-precision map; the database stores feature information of ground features extracted based on historical data, but this feature information has not yet been used to establish a new high-precision map. Therefore, the change status can include whether feature information of ground features has been extracted. In practical applications, multiple change statuses can be set according to business needs; this embodiment does not limit this. Different update processes can be configured for different states of change, so that newly collected data can be updated according to the first data of different states of change, thereby improving update efficiency.

[0095] Based on this, the identification of changing states can be divided into three categories:

[0096] The identification of changes involves determining whether any changes have occurred in the target geographic area based on the new trajectory image association data and the historical trajectory image association data. For example, matching new and historical image trajectory points are obtained from the new and historical trajectory image association data. The new images associated with the new trajectory points and the historical images associated with the historical trajectory points are compared, and the comparison results determine whether any changes have occurred in the location of the new trajectory points.

[0097] The identification of historical data involves determining the location of newly acquired data and whether there is historical data available. For example, for new image trajectory points included in new trajectory image association data, the system searches the historical trajectory image association data for matching historical image trajectory points.

[0098] The identification of whether or not there is a base map is to determine the location of newly collected data and whether there are corresponding historical features in historically collected data.

[0099] In some examples, the above three identification processes can be implemented using three independent services. Each newly collected data segment can be used as input for the three services, which are then called for identification. Each service can configure an indicator indicating the change status of each newly collected data segment based on the change status of the newly collected data segment determined in each of the determination steps.

[0100] In other examples, the above identification steps may be executed in a specific order, which can be configured as needed in practical applications. This embodiment does not limit this. For example, the identification of whether there is a history can be performed first. If there is no historical image trajectory point with a matching location, it can be determined that the location is a newly acquired geographical location. Since no high-precision map has been established for this geographical location, it can also be determined at the same time that there is no corresponding historical image and historical feature information for this location.

[0101] For example, the identification of the presence or absence of a background map and the presence or absence of historical data can be handled automatically; the identification of whether there are changes can also be automated, and review tasks can be generated to assist manual completion. As another example, a single user performing an entire project's tasks is inefficient and of low quality. This embodiment employs segmented processing, dividing the large volume of newly collected data into smaller, more granular segments. A review task can be generated for each segment and provided to the client for display. Client-side operators can perform tasks through the client and upload the results, i.e., input change status information through the client. For example, such as... Figure 2B The diagram shown is a schematic representation of a client according to an exemplary embodiment of this specification. It shows a segment of newly acquired data and a segment of historical acquired data matching its location. By reviewing the new and historical images, operators can determine whether there are any changed features. Operators can upload the determination results through the client, for example, by adding a marker if it is determined that there are changed features.

[0102] In some examples, the client may also include functionality for users to input other information. For instance, users can input information about changed feature elements, such as their category. As an example, the client can implement an input control for inputting information about changed feature elements, allowing users to perform input operations. For instance, the input control could be a graphic box control, such as a rectangle control. Users can define a graphic box in the new image; the position of this graphic box represents the pixel position of the changed feature element in the new image. The client can obtain the pixel position information of the graphic box through this control. Alternatively, the input control could be a category selection control or a text editing control, used for users to input the category information of the changed feature element. For example, as mentioned earlier, feature elements include many categories, such as lane lines, signs, and so on.

[0103] Based on the aforementioned identification of whether or not changes have occurred, the change status of each segment of newly collected data can be obtained. Next, data classification and aggregation can be performed. Different update processing flows correspond to different change statuses, and each update processing flow can handle multiple change statuses.

[0104] As an example, optionally, each change state corresponds to a processing flow. In practical applications, various change states and their corresponding update processing flows can be determined as needed.

[0105] For example, the changed state includes the absence of historical trajectory points in the historical trajectory data that correspond to the new image trajectory points in the data associated with the new trajectory image, and the update process includes establishing a high-precision map using the newly collected data corresponding to the new image trajectory points.

[0106] As an example, in practical applications, databases are established to store historical data of established high-precision maps, such as historical trajectory points, historical images, historical point clouds, or historical feature information. In this embodiment, historical trajectory points carry geographical locations, allowing us to search for historical trajectory points in the collected data that correspond to the geographical location of the first data. If the change status is no historical trajectory point, it can be determined that a high-precision map has not been established at the location of the new image trajectory point. Based on this, the corresponding update process can be to establish a high-precision map corresponding to the geographical location based on the newly collected data. If the change status is that historical trajectory points exist, it indicates that data has been collected at that geographical location. The subsequent update process can be further determined by combining the search results of historical images and historical feature information.

[0107] For example, if the geographical location of a newly collected data segment is identified as the newly collected location, and there are no corresponding historical image trajectory points and historical images in the historical data, the corresponding update process can be to build a high-precision map corresponding to the geographical location based on the newly collected data.

[0108] For example, if a newly acquired data set contains historical trajectory image association data that matches the location of the newly acquired image association data (i.e., the same location has both newly acquired and historical images), the change state of the new image relative to the historical image can be determined using the newly acquired and historical images, i.e., whether any ground features have changed. For instance, the change state may include the newly acquired image not changing relative to the historical image, in which case the update process includes storing the newly acquired data for that location. And / or, the change state may include the presence of changed ground features, in which case the update process includes updating the high-precision map based on the newly acquired data for that location.

[0109] For example, the changed state includes situations where the location of a new trajectory point in the newly acquired data matches that of a historical trajectory point, but there is no historical feature information at that location. The update process includes establishing a high-precision map based on the historical feature information and the newly acquired data corresponding to that location after the historical feature information for that location is established; alternatively, it can interrupt the existing feature information generation process and establish a high-precision map corresponding to that geographical location based on the newly acquired data. In this embodiment, the geographical location of the new trajectory point in the newly acquired data has historical trajectory points matching its location in the historical acquired data, but there is no corresponding historical feature information. This situation exists because multiple processes need to be performed before establishing a high-precision map, such as projecting image data and point cloud data, generating and storing feature information of ground features from point clouds and / or image data, etc. The process of generating feature information usually takes a certain amount of time, and the absence of historical feature information also indicates that the high-precision map has not yet been established. Based on this, the above update process can reduce processing costs and improve the update efficiency of high-precision maps.

[0110] like Figure 2CThe diagram illustrates another method for processing newly acquired sub-data according to an exemplary embodiment of this specification. In the diagram, the original data represents the newly acquired data; changed data represents the newly acquired data corresponding to the locations where ground features have changed; and unchanged data represents the newly acquired data corresponding to the locations where ground features have changed. In cases of change, corresponding processing flows can be executed, such as the cloud upload, data segmentation, data alignment, data recognition, and manual operations shown in the diagram to create a high-precision map. In cases of no change, only the data acquisition time needs to be updated in the database; no other processes are required. This significantly improves the efficiency of high-precision map data processing and data production, reduces production costs, and shortens the delivery cycle.

[0111] Corresponding to the embodiments of the aforementioned method for detecting changes in ground features, this specification also provides embodiments of a device for detecting changes in ground features and the computer equipment used therein.

[0112] The embodiments of the feature change detection device described in this specification can be applied to computer equipment, such as servers or terminal devices. The device embodiments can be implemented through software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it is formed by its processor reading the corresponding computer program instructions from non-volatile memory into memory and executing them. From a hardware perspective, such as... Figure 3 The diagram shown is a hardware structure diagram of the computer equipment housing the feature change detection device in this specification. (Except for...) Figure 3 In addition to the processor 310, memory 330, network interface 320, and non-volatile memory 340 shown, the computer device where the feature change detection device 331 is located in the embodiment may also include other hardware depending on the actual function of the computer device, which will not be described in detail here.

[0113] like Figure 4 As shown, Figure 4 This is a block diagram illustrating a change detection device for ground features according to an exemplary embodiment of this specification, the device comprising:

[0114] New data acquisition module 41 is used to: acquire newly collected data of the target geographic area, the newly collected data including new trajectory data and new image data;

[0115] The association module 42 is used to: associate the new trajectory data with the new image data to obtain new trajectory image association data;

[0116] Historical data acquisition module 43 is used to: acquire historical trajectory image association data of the target geographic area;

[0117] The change determination module 44 is used to: determine whether there are any changed geographical features in the target geographic area based on the new trajectory image association data and the historical trajectory image association data.

[0118] In some examples, the associated module is also used for:

[0119] New image trajectory points are added to the new trajectory data, and the new images in the new image data are associated with the new image trajectory points to obtain new trajectory image association data.

[0120] In some examples, the device further includes a thinning module for:

[0121] The trajectory points in the new trajectory data are thinned out according to a set rule to obtain the thinned new trajectory data;

[0122] The associated module is also used for:

[0123] New image trajectory points are added to the thinned new trajectory data, and the new images in the new image data are associated with the new image trajectory points to obtain new trajectory image association data.

[0124] In some examples, the associated module is also used for:

[0125] The shooting location of the new image, including the new image data, is added to the thinned new trajectory data as the new image trajectory point corresponding to the new image;

[0126] The new image is associated with its corresponding new image trajectory points to obtain new trajectory image association data.

[0127] In some examples, the associated module is also used for:

[0128] Based on the new trajectory points included in the thinned new trajectory data, interpolation processing is performed to obtain the positions of the new image trajectory points;

[0129] The new image data includes matching the shooting location of the new image with the location of the trajectory point of the new image, and then associating the new image with the matching location with the trajectory point of the new image to obtain new trajectory image association data.

[0130] In some examples, the device further includes a segmentation processing module for:

[0131] The newly collected data for the target geographic area is segmented and processed to obtain segmented and stored new data.

[0132] In some examples, the segmentation processing module is also used for:

[0133] Obtain a pre-made road surface bounding box, which represents the three-dimensional spatial information of the road. The road surface bounding box is made based on road data from an existing high-precision map, which is related to the historical trajectory image association data.

[0134] The road surface bounding box is used to determine the road to which the new trajectory data in the newly collected data belongs;

[0135] The newly collected data for the target geographic area is segmented based on the road to which the new trajectory data belongs, resulting in segmented and stored new data.

[0136] In some examples, the change determination module is also used for:

[0137] Obtain new image trajectory points and historical image trajectory points with matching positions from the new trajectory image association data and the historical trajectory image association data;

[0138] Compare the new image associated with the new image trajectory points with the historical image associated with the historical image trajectory points;

[0139] Based on the comparison results, determine whether there are any changed ground features at the location of the new image trajectory points.

[0140] In some examples, the change determination module is also used for:

[0141] If it is determined that the location of the new image trajectory point has changed ground features, the high-precision map needs to be updated to determine the location of the new image trajectory point;

[0142] If it is determined that the location of the new image trajectory point has not changed any ground features, the new image trajectory point and its associated new image are stored.

[0143] In some examples, the device further includes a lookup module for:

[0144] For the new image trajectory points included in the new trajectory image association data, search the historical trajectory image association data for historical image trajectory points with matching locations; and / or,

[0145] The search module is also used for:

[0146] Historical feature data is acquired, wherein the historical feature data includes historical feature information generated using historical image data corresponding to historical image trajectory points in the historical trajectory image association data, and / or historical point cloud data corresponding to the historical image data;

[0147] For the new image trajectory points included in the new trajectory image association data, search the historical feature data for historical feature information that matches the location.

[0148] The specific implementation process of the functions and roles of each module in the aforementioned feature change detection device is detailed in the implementation process of the corresponding steps in the aforementioned feature change detection method, and will not be repeated here.

[0149] Accordingly, embodiments of this specification also provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the aforementioned method embodiment for discovering changes in ground features.

[0150] Accordingly, embodiments of this specification also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method embodiment for discovering changes in ground features.

[0151] Accordingly, embodiments of this specification also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method embodiment for discovering changes in ground features.

[0152] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of the solution in this specification according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0153] The method for discovering changes in ground features is applied to one or more electronic devices. The electronic device is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions. The hardware of the electronic device includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0154] The electronic device can be any electronic product that can interact with the user, such as a personal computer, tablet computer, smartphone, personal digital assistant (PDA), game console, interactive network television (IPTV), smart wearable device, etc.

[0155] The electronic device may also include network devices and / or user devices. The network devices include, but are not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of hosts or network servers.

[0156] The networks in which the electronic devices are located include, but are not limited to, the Internet, wide area networks, metropolitan area networks, local area networks, and virtual private networks (VPNs).

[0157] The specific implementation process of the above embodiments can be referred to the description of the foregoing embodiments, and will not be repeated here.

[0158] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0159] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this application.

[0160] The terms "specific example" or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with the embodiments or examples, which are included in at least one embodiment or example of this specification. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0161] Other embodiments of this specification will readily occur to those skilled in the art upon consideration of the specification and practice of the invention claimed herein. This specification is intended to cover any variations, uses, or adaptations that follow the general principles of this specification and include common knowledge or customary techniques in the art not claimed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this specification are indicated by the following claims.

[0162] It should be understood that this specification is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this specification is limited only by the appended claims.

[0163] The above description is merely a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

Claims

1. A method for detecting changes in ground features, comprising: Acquire newly collected data for the target geographic area, including new trajectory data and new image data; Based on the time information and / or geographical location information of trajectory points in the new trajectory data, and the time information and / or shooting location of images in the new image data, the new trajectory data and the new image data are associated to obtain new trajectory image association data; Obtain historical trajectory image association data of the target geographic area; Based on the new trajectory image association data and the historical trajectory image association data, determine whether there are any changed geographical features in the target geographic area; The newly collected data of the target geographic area is segmented and stored after the road to which the new trajectory data belongs is determined by the road surface frame. The road surface frame is created in advance using the three-dimensional road data in the generated high-precision map.

2. The method according to claim 1, wherein, The step of associating the new trajectory data with the new image data to obtain new trajectory image association data includes: New image trajectory points are added to the new trajectory data, and the new images in the new image data are associated with the new image trajectory points to obtain new trajectory image association data.

3. The method according to claim 2, wherein, The method further includes: The trajectory points in the new trajectory data are thinned out according to a set rule to obtain the thinned new trajectory data; The step of adding new image trajectory points to the new trajectory data and associating the new images in the new image data with the new image trajectory points to obtain new trajectory image association data includes: New image trajectory points are added to the thinned new trajectory data, and the new images in the new image data are associated with the new image trajectory points to obtain new trajectory image association data.

4. The method according to claim 3, wherein, The step of adding new image trajectory points to the thinned new trajectory data and associating the new images in the new image data with the new image trajectory points to obtain new trajectory image association data includes: The shooting location of the new image, including the new image data, is added to the thinned new trajectory data as the new image trajectory point corresponding to the new image; The new image is associated with its corresponding new image trajectory points to obtain new trajectory image association data.

5. The method according to claim 3, wherein, The step of adding new image trajectory points to the thinned new trajectory data and associating the new images in the new image data with the new image trajectory points to obtain new trajectory image association data includes: Based on the new trajectory points included in the thinned new trajectory data, interpolation processing is performed to obtain the positions of the new image trajectory points; The new image data includes matching the shooting location of the new image with the location of the trajectory point of the new image, and then associating the new image with the matching location with the trajectory point of the new image to obtain new trajectory image association data.

6. The method according to any one of claims 1 to 5, wherein, The newly collected data, which is stored in segments, is obtained in the following way: Obtain a pre-made road surface bounding box, which represents the three-dimensional spatial information of the road, and the generated high-precision map is related to the historical trajectory image association data; The road surface bounding box is used to determine the road to which the new trajectory data in the newly collected data belongs; The newly collected data for the target geographic area is segmented based on the road to which the new trajectory data belongs, resulting in segmented and stored new data.

7. The method according to any one of claims 1 to 5, wherein, The step of determining whether there are changed geographic features in the target geographic area based on the new trajectory image association data and the historical trajectory image association data includes: Obtain new image trajectory points and historical image trajectory points with matching positions from the new trajectory image association data and the historical trajectory image association data; Compare the new image associated with the new image trajectory points with the historical image associated with the historical image trajectory points; Based on the comparison results, determine whether there are any changed ground features at the location of the new image trajectory points.

8. The method according to claim 1, further comprising: For the new image trajectory points included in the new trajectory image association data, search in the historical trajectory image association data for historical image trajectory points with matching locations; And / or, The method further includes: Historical feature data is acquired, wherein the historical feature data includes historical feature information generated using historical image data corresponding to historical image trajectory points in the historical trajectory image association data, and / or historical point cloud data corresponding to the historical image data; For the new image trajectory points included in the new trajectory image association data, search the historical feature data for historical feature information that matches the location.

9. A device for detecting changes in ground features, the device comprising: The new data acquisition module is used to: acquire newly collected data of the target geographic area, the newly collected data including new trajectory data and new image data; The association module is used to: associate the new trajectory data with the new image data based on the time information and / or geographical location information of the trajectory points in the new trajectory data, and the time information and / or shooting location of the images in the new image data, to obtain new trajectory image association data; The historical data acquisition module is used to: acquire historical trajectory image association data of the target geographic area; The change determination module is used to: determine whether there are any changed geographical features in the target geographic area based on the new trajectory image association data and the historical trajectory image association data; The newly collected data of the target geographic area is segmented and stored after the road to which the new trajectory data belongs is determined by the road surface frame. The road surface frame is created in advance using the three-dimensional road data in the generated high-precision map.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 8.

Images