Vector-based Map Processing Method and Device

Abstract

The embodiments of this specification provide a vector-based map processing method and apparatus. Among them, the method includes: matching the initial vector and the updated vector of the target map according to a preset matching rule to determine a vector matching pair; determining the pose adjustment parameter of the updated vector and the relocalization parameter of each vector point in the updated vector according to the vector matching pair; adjusting the updated vector according to the pose adjustment parameter and the relocalization parameter to obtain an adjusted updated vector; determining the confidence level of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pair and the association relationship between the initial vector and the adjusted updated vector; and determining whether the target map has changed according to the vector matching pair, the pose adjustment parameter, and / or the confidence level of the updated vector.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of computer technology, and particularly to a vector-based map processing method. Background Art

[0002] High-precision maps accurately describe the position, attributes, geometry and other information of various road elements (such as bus stops, road signs, etc.) in the real world, and are one of the data sources for autonomous driving technology. Most autonomous driving or assisted driving vehicles on the market rely on high-precision maps to help complete tasks such as perception, positioning, and planning. This requires high-precision maps to accurately and real-time reflect the actual state of road elements. That is, it is necessary to accurately discover the currency changes in high-precision maps, so as to update the high-precision maps in a timely manner.

[0003] Currently, for the currency changes in high-precision maps, it is generally a manual carpet search to determine whether a certain road element has changed. Subsequently, after manually determining the location of the change of the road element, on-site measurement, data collection, etc. are carried out to update the high-precision map; in this way, the currency changes of road elements in high-precision maps cannot be obtained in a timely manner, and it is realized manually, with a long processing cycle, high cost, and poor accuracy. Summary of the Invention

[0004] In view of this, the embodiments of this specification provide a vector-based map processing method. One or more embodiments of this specification also relate to a vector-based map processing device, a computing device, a computer-readable storage medium, and a computer program to solve the technical defects existing in the prior art.

[0005] According to the first aspect of the embodiments of this specification, a vector-based map processing method is provided, including:

[0006] Matching the initial vector and the update vector of the target map according to a preset matching rule to determine a vector matching pair;

[0007] According to the vector matching pair, determining the pose adjustment parameter of the update vector and the repositioning parameter of each vector point in the update vector;

[0008] Adjusting the update vector according to the pose adjustment parameter and the repositioning parameter to obtain an adjusted update vector;

[0009] Determining the confidence level of the update vector according to the matching distance between the initial vector and the update vector in the vector matching pair, and the correlation between the initial vector and the adjusted update vector;

[0010] Determine whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vectors.

[0011] According to a second aspect of the embodiments of the present specification, a vector-based map processing apparatus is provided, including:

[0012] A vector matching module, configured to match the initial vectors and the updated vectors of the target map according to a preset matching rule to determine vector matching pairs;

[0013] A parameter calculation module, configured to determine the pose adjustment parameters of the updated vectors and the relocalization parameters of each vector point in the updated vectors according to the vector matching pairs;

[0014] A vector adjustment module, configured to adjust the updated vectors according to the pose adjustment parameters and the relocalization parameters to obtain adjusted updated vectors;

[0015] A confidence level determination module, configured to determine the confidence level of the updated vectors according to the matching distance between the initial vectors and the updated vectors in the vector matching pairs and the correlation between the initial vectors and the adjusted updated vectors;

[0016] A change determination module, configured to determine whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vectors.

[0017] According to a third aspect of the embodiments of the present specification, a computing device is provided, including:

[0018] A memory and a processor;

[0019] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the above-mentioned vector-based map processing method are implemented.

[0020] According to a fourth aspect of the embodiments of the present specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned vector-based map processing method are implemented.

[0021] According to a fifth aspect of the embodiments of the present specification, a computer program is provided, wherein when the computer program is executed on a computer, the computer is made to execute the steps of the above-mentioned vector-based map processing method.

[0022] One embodiment of this specification implements a vector-based map processing method and apparatus. The method includes matching an initial vector and an updated vector of a target map according to a preset matching rule to determine a vector matching pair; determining a pose adjustment parameter of the updated vector and a relocalization parameter of each vector point in the updated vector according to the vector matching pair; adjusting the updated vector according to the pose adjustment parameter and the relocalization parameter to obtain an adjusted updated vector; determining a confidence level of the updated vector according to a matching distance between the initial vector and the updated vector in the vector matching pair and an association relationship between the initial vector and the adjusted updated vector; and determining whether the target map has changed according to the vector matching pair, the pose adjustment parameter, and / or the confidence level of the updated vector. Specifically, the vector-based map processing method can detect changes in the real world in a timely and accurate manner based on the initial vector data and the updated vector data of the target map, maintain the timeliness of the data in the target map, and greatly shorten the processing cycle of the currency changes of road elements in the high-precision map and reduce costs by using such an automated detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flowchart of a vector-based map processing method provided by an embodiment of this specification;

[0024] Figure 2 is a flowchart of obtaining a vector matching pair in a vector-based map processing method provided by an embodiment of this specification;

[0025] Figure 3 is a flowchart of updating the vector point coordinates of an updated vector in a vector-based map processing method provided by an embodiment of this specification;

[0026] Figure 4 is a flowchart of obtaining the confidence level of an updated vector in a vector-based map processing method provided by an embodiment of this specification;

[0027] Figure 5 is a judgment flowchart of determining whether a target map has changed in a vector-based map processing method provided by an embodiment of this specification;

[0028] Figure 6 is a schematic structural diagram of a vector-based map processing apparatus provided by an embodiment of this specification;

[0029] Figure 7 is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] In the following description, numerous specific details are set forth to provide a thorough understanding of this specification. However, this specification can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the connotation of this specification. Therefore, this specification is not limited by the specific implementations disclosed below.

[0031] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms "a", "the", and "said" used in one or more embodiments of 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" used in one or more embodiments of this specification refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0032] It should be understood that although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0033] First, the noun terms related to one or more embodiments of this specification are explained.

[0034] High-precision map: A high-precision map for autonomous driving.

[0035] Differential change discovery: Comparing the differences between a single collected map and the features on the real world.

[0036] Vector: Various high-precision map elements composed of points containing information such as three-dimensional coordinates and attributes.

[0037] Base vector: Vector elements that have been made into a map and are to be updated.

[0038] Updated vector: Map vectors currently being made and not yet formed into a map.

[0039] In specific implementation, for map updates, the areas where the features have changed can be measured on-site, data can be collected, and then the map database can be updated. By this method, the current changes in the map can be accurately discovered. However, on the one hand, a carpet-like investigation is required, with a long time cycle and high cost; on the other hand, it is necessary to manually determine the positions where the map elements have changed, and the accuracy is low.

[0040] It is also possible to update the surface and features of a high-precision map at a relatively small spatial scale based on high-resolution remote sensing image classification. This method can efficiently update features with obvious geometric features, but is limited by image resolution, image classification technology, etc., and cannot accurately identify road elements with unclear geometric features such as poles, signs, and lane lines in high-precision maps.

[0041] In addition, a crowdsourcing map update solution based on vision technology can be adopted to update high-precision maps using a large amount of data, with relatively low costs. However, this solution requires a large amount of data for model training and has a relatively low accuracy in scenarios with a small amount of data.

[0042] Based on this, in this specification, a vector-based map processing method is provided. This specification also relates to a vector-based map processing device, a computing device, a computer-readable storage medium, and a computer program, which will be described in detail one by one in the following embodiments.

[0043] See Figure 1 , Figure 1 shows a flowchart of a vector-based map processing method provided according to an embodiment of this specification, which specifically includes the following steps.

[0044] Step 102: Match the initial vector and the update vector of the target map according to a preset matching rule to determine a vector matching pair.

[0045] Among them, the preset matching rule can be set according to actual applications, and the embodiments of this specification do not make limitations; the target map can be understood as a map of any type and any range, such as world maps, national maps, city maps, high-precision maps of any type, etc.; the initial vector of the target map can be understood as a base vector, that is, a vector element that has been made into a map and is to be updated. Among them, a vector can be understood as various road elements (poles, signs, lane lines, etc.) of a high-precision map composed of points containing three-dimensional coordinates, attributes, and other information; the update vector of the target map can be understood as a map vector that is currently being made and has not been formed into a map.

[0046] Specifically, after obtaining the target map and the initial vector and update vector of the target map, the vector points in the initial vector can be matched with the vector points in the update vector through a preset matching rule to accurately obtain the vector matching pair of the initial vector and the update vector. The specific implementation method is as follows:

[0047] The matching of the initial vector and the update vector of the target map according to the preset matching rule to determine a vector matching pair includes:

[0048] Obtain the initial vector and the updated vector of the target map, and respectively determine the geometric types of the road elements corresponding to the vector points in the initial vector and the updated vector;

[0049] According to the geometric types, respectively fit the initial vector and the updated vector through a preset fitting algorithm to obtain the road geometric elements of the initial vector and the updated vector;

[0050] Determine vector matching pairs according to the initial vector, the updated vector, the road geometric elements of the initial vector, and the road geometric elements of the updated vector.

[0051] Among them, road elements include but are not limited to poles, signs, lane lines, etc. The geometric types of road elements can be understood as the types of poles, signs, lane lines, etc. For example, the type of a pole is a point, the type of a sign is a surface, and the type of a lane line is a straight line, etc.

[0052] In practical applications, both the initial vector and the updated vector include multiple vector points. Each vector point contains information such as three-dimensional coordinates and attributes. By fitting these vector points according to a preset fitting algorithm, the road elements in the initial vector and the updated vector can be obtained.

[0053] Specifically, obtain the target map (such as a high-precision map for determining whether there is a change in road elements), as well as the base vector and the updated vector of the target map; determine the geometric types of the road elements corresponding to the vector points in the initial vector (such as straight lines, points, planes, etc.), and the geometric types of the road elements corresponding to the vector points in the updated vector (such as straight lines, points, planes, etc.); respectively fit the initial vector and the updated vector through a preset fitting algorithm according to each geometric type. Among them, when the geometric types are different, the preset fitting algorithms used for fitting the initial vector and the updated vector can also be different. For example, when the geometric type is a straight line, the least squares method can be used to fit the straight line; and after fitting the initial vector and the updated vector through a preset fitting algorithm according to the geometric type, respectively obtain the road geometric elements of the initial vector and the updated vector (such as line elements, surface elements, or point elements, etc.).

[0054] Then, determine the vector matching pairs of the initial vector and the updated vector according to the initial vector, the updated vector, the road geometric elements of the initial vector, and the road geometric elements of the updated vector.

[0055] During specific implementation, when obtaining the road geometric elements of the initial vector and the updated vector, it is necessary to first obtain the initial road geometric elements of the initial vector and the updated vector through a fitting algorithm, and then perform other processing on the initial road geometric elements of the initial vector and the updated vector to determine the accurate road geometric elements of the final initial vector and the updated vector. The specific method is as follows:

[0056] Fitting the initial vector and the updated vector respectively through a preset fitting algorithm according to the geometric type to obtain the road geometric elements of the initial vector and the updated vector, including:

[0057] Determine the fitting algorithms corresponding to the initial vector and the updated vector according to the geometric type;

[0058] Obtain the initial road geometric elements of the initial vector according to the fitting algorithm corresponding to the initial vector;

[0059] Obtain the initial road geometric elements of the updated vector according to the fitting algorithm corresponding to the updated vector;

[0060] Adjust the initial road geometric elements of the initial vector and the initial road geometric elements of the updated vector according to a preset calculation rule to obtain the road geometric elements of the initial vector and the updated vector.

[0061] Among them, the fitting algorithm includes but is not limited to the least squares method.

[0062] In specific implementation, the fitting algorithms corresponding to different geometric types may be different. For example, one fitting algorithm corresponds to a plane, and another fitting algorithm corresponds to a straight line, etc. First, determine the fitting algorithm corresponding to the initial vector according to the geometric type, and the fitting algorithm corresponding to the updated vector. Then, obtain the initial road geometric elements of the initial vector according to the fitting algorithm corresponding to the initial vector, and obtain the initial road geometric elements of the updated vector according to the fitting algorithm corresponding to the updated vector. Subsequently, adjust the initial road geometric elements of the initial vector and the initial road geometric elements of the updated vector according to a preset calculation rule to obtain the road geometric elements of the initial vector and the updated vector to ensure the accuracy of the road geometric elements of the initial vector and the updated vector.

[0063] Specifically, the adjusting the initial road geometric elements of the initial vector and the initial road geometric elements of the updated vector according to a preset calculation rule to obtain the road geometric elements of the initial vector and the updated vector includes:

[0064] Calculate the first fitting residual of the vector points in the initial vector according to the initial road geometric elements of the initial vector, and calculate the root mean square, mean, and standard deviation of the first fitting residual according to the first fitting residual;

[0065] Calculate the second fitting residual of the vector points in the updated vector according to the initial road geometric elements of the updated vector, and calculate the root mean square, mean, and standard deviation of the second fitting residual according to the second fitting residual;

[0066] Adjust the initial road geometric elements of the initial vector according to the root mean square, mean, and standard deviation of the first fitting residual to obtain the road geometric elements of the initial vector;

[0067] Adjust the initial road geometric elements of the updated vector according to the root mean square, mean, and standard deviation of the second fitting residual to obtain the road geometric elements of the updated vector.

[0068] In practical applications, after respectively determining the initial road geometric elements of the initial vector and the updated vector, calculate the first fitting residual of each vector point of the initial vector and the second fitting residual of each vector point of the updated vector according to the initial road geometric elements of the initial vector and the updated vector. Then calculate the root mean square, mean, and standard deviation of the first fitting residual based on the first fitting residual of each vector point of the initial vector, and at the same time calculate the root mean square, mean, and standard deviation of the second fitting residual based on the second fitting residual of each vector point of the updated vector; subsequently, adjust the initial road geometric elements of the initial vector according to the root mean square, mean, and standard deviation of the first fitting residual to obtain the road geometric elements of the initial vector, and at the same time adjust the initial road geometric elements of the updated vector according to the root mean square, mean, and standard deviation of the second fitting residual to obtain the road geometric elements of the updated vector.

[0069] Among them, the adjustment of the initial road geometric elements of the initial vector can be understood as marking the outlier points that deviate from twice the standard deviation of the mean of the first fitting residual according to the root mean square, mean, and standard deviation of the first fitting residual, and removing the marked outlier points when the root mean square of the first fitting residual does not meet the preset requirements (such as a preset distance threshold, 3 cm, 5 cm, etc.). Repeat to obtain the initial road geometric elements of the initial vector through the fitting algorithm, and adjust the initial road geometric elements of the initial vector according to the above preset calculation rules until the root mean square of the first fitting residual meets the preset requirements and then end, so as to determine the road geometric elements of the initial vector; similarly, the determination of the road geometric elements of the updated vector can refer to the determination method of the road geometric elements of the initial vector, which will not be elaborated here.

[0070] Taking the geometric types of road elements corresponding to vector points in the initial vector and the updated vector including straight lines and planes as an example, when fitting the straight lines and planes in the initial vector and the updated vector through a fitting algorithm, the calculated fitting residual refers to (taking the updated vector as an example) that n vector points fit a straight line L, and then the residuals between the n points and the straight line L are calculated. That is, substituting the vector point coordinates (x0, y0) into the straight line equation of L, (x0, y1) will be obtained, and r = y1 - y0, where r is the fitting residual. Since there are n points, n rs will be obtained, which is defined as the set R. Then, the root mean square error of R is calculated, that is, the fitting residual rms. At the same time, the mean and standard deviation of R are calculated, and then the outlier points are removed to obtain the accurate road geometric elements after adjusting the initial vector and the updated vector. Among them, the purpose of deleting outlier points is to remove abnormal gross errors as much as possible to ensure the accuracy of fitting. Of course, if too many outliers are removed, then the vector fitting will be marked as failed.

[0071] After determining the road geometric elements of the initial vector and the updated vector, the vector matching pairs between the initial vector and the updated vector can be calculated quickly and accurately. The specific implementation method is as follows:

[0072] Determining the vector matching pairs according to the initial vector, the updated vector, the road geometric elements of the initial vector, and the road geometric elements of the updated vector includes:

[0073] Calculating the distance error between the vector points of the updated vector and the road geometric elements of the initial vector, and determining the initial vector matching pairs according to the distance error;

[0074] Determining the attribute information of the vector points of the updated vector in the initial vector matching pairs and the attribute information of the corresponding vector points of the initial vector;

[0075] When the attribute information of the two is the same, calculating the included angle between the road geometric elements of the updated vector and the road geometric elements of the initial vector in the initial vector matching pairs, and determining the vector matching pairs according to the included angle.

[0076] Specifically, calculate the distance error between each vector point in the updated vector and each road geometric element in the initial vector, and delete the bottom vector with a distance error greater than the preset distance threshold to screen out the road geometric elements of the initial vector corresponding to each vector point in the updated vector, and record the vector pairs that meet the requirements. For example, if the distance errors between the four vector points in the updated vector and the surface element in the road geometric elements of the initial vector are all less than or equal to the preset distance threshold, it means that these four vector points and the surface element in the road geometric elements of the initial vector form the initial vector matching pairs.

[0077] Then determine the attribute information of each vector point in the initial vector matching pair. When the attribute information of the two vector points is consistent, calculate the angle between the road geometry element of the updated vector in the initial vector matching pair and the road geometry element of the initial vector. When the angle is less than or equal to the preset angle threshold, record the vector pair that meets the requirements and use it as the final vector matching pair.

[0078] In practical applications, when calculating the distance error, for line elements (line elements in road geometric elements), the distance error between each vector point in the update vector and each line element in the initial vector is calculated one by one, the initial vectors with distance errors greater than the preset distance threshold are eliminated, and the vector pairs that meet the requirements are recorded; for surface elements (surface elements in road geometric elements), the distance error between each vector point in the update vector and each surface element in the initial vector is calculated one by one, the initial vectors with distance errors greater than the preset distance threshold are eliminated, and the vector pairs that meet the requirements are recorded.

[0079] When calculating the distance error between the vector points of the update vector and the road geometric elements of the initial vector, the base vector has been fitted into a straight line or plane, while the update vector is a bunch of vector scattered points. Taking the straight line fitted by the base vector as an example, suppose there are n vector points of the update vector, calculate the distance between the n points and the straight line fitted by the base vector one by one, and obtain the set R. Calculate the rms (root mean square) of R, and judge whether there is a corresponding relationship between the two straight lines based on the rms. In actual application, there are certain differences in the calculation methods of different elements, which are determined according to the actual application.

[0080] It is determined whether the attributes of the vector points of the initial vector and the vector points of the update vector are consistent in the vector pair that meets the requirements. If not, the updated vector and all the vector points in the initial vector are continuously used to fit line features or surface features by the least squares method. If they are consistent, the angle between the road geometric features of the update vector and the road geometric features of the initial vector in the initial vector matching pair is calculated.

[0081] When calculating the angle, for line elements (line elements in road geometry elements), the angle between each line element in the update vector and each line element in the initial vector is calculated one by one in the initial vector matching pair. If the angle is less than or equal to the preset angle threshold, the vector pair that meets the requirements is recorded; for surface elements (surface elements in road geometry elements), the angle between the normal vector of each surface element in the update vector and each surface element in the initial vector is calculated one by one. If the angle is less than or equal to the preset angle threshold, the vector pair that meets the requirements is recorded and used as the final vector matching pair. If the angle is greater than the preset angle threshold, it is considered that the matching relationship does not exist. For example, if the angle between two straight lines exceeds a certain threshold, it can be considered that the two straight lines are not corresponding.

[0082] Taking the straight line element of the ground lane line as an example, if the straight line equation of the fitted base vector is: Ax + By + C = 0;

[0083] The vector point of the updated vector is: p(x0, y0), then the distance from the point to the straight line is obtained by calculating with Formula 1:

[0084]

[0085] where A, B, and C are all undetermined parameters of the straight line equation; X and Y are the X coordinate and Y coordinate in the straight line equation respectively.

[0086] Taking the surface element of the card as an example, if the plane equation of the fitted base vector is: Ax + By + Cz + D = 0;

[0087] The vector point of the updated vector is p(x0, y0, z0), then the distance from the point to the plane is obtained by calculating with Formula 2:

[0088]

[0089] For the surface element, the vector point of the base vector is generally composed of n corner points. Taking the rectangular points as an example, then n = 4;

[0090] The center of gravity point of the current plane can be calculated from the 4 corner points, see Formula 3:

[0091]

[0092] Then the updated vector will also have corresponding n + 1 points. After aligning the order, calculate the distance between points respectively, see Formula 4:

[0093] Formula 4:

[0094]

[0095] Then 5 distances d are obtained, and then calculate the rms (root mean square) of the 5 distances d, which is the final distance from the point to the plane.

[0096] In specific use, the calculation method of Formula 2 is used for preliminary screening, and then the calculation methods of Formula 3 and Formula 4 are used as the final basis.

[0097] See Figure 2 , Figure 2 shows a flowchart of obtaining vector matching pairs in a vector-based map processing method provided according to an embodiment of this specification, which specifically includes the following steps.

[0098] Step 202: Determine the element types of the base vector and the updated vector respectively.

[0099] The base vector may be understood as the initial vector of the above embodiment, and the element type may be understood as the base vector of the above embodiment and the geometric type of the road element corresponding to the vector point in the update vector.

[0100] Take the example that the feature type includes line features and area features.

[0101] Step 204: For the base vector and the update vector, all the vector points are used to obtain the line elements of the base vector and the update vector respectively through least square fitting.

[0102] Step 206: For the base vector and the update vector, all vector points are used to obtain the surface elements of the base vector and the update vector respectively through least square fitting.

[0103] Step 208: Perform linear anti-error fitting, attribute consistency check, angle consistency check, and linear distance check calculation on the line elements.

[0104] Specifically, straight line robust fitting: for the line element of the base vector or the update vector, calculate the fitting residuals of each vector point in the base vector and the update vector, calculate the rms, mean and standard deviation of the fitting residuals, and mark the outliers that deviate from the mean of the fitting residuals by 2 times the standard deviation; when the rms of the fitting residuals does not meet the preset requirements, remove the outliers, repeat steps 204 to 206 until the rms of the fitting residuals meet the preset requirements, and determine the final line element of the base vector or the update vector.

[0105] Straight-line distance check: Calculate the distance error between each vector point in the update vector and the line feature in the base vector one by one, eliminate the base vectors whose distance error is greater than the threshold, and record the vector pairs that meet the requirements.

[0106] Attribute consistency check: Determine whether the attributes of the update vector and the base vector in the vector pair that meets the requirements determined by the straight-line distance check are consistent. If they are inconsistent, it indicates that the vector pair does not match. If they are consistent, perform an angle consistency check on them to obtain the final vector pair that meets the requirements.

[0107] Angle consistency check: Calculate the angle between each line feature in the update vector and the line feature in the base vector in the vector pairs that meet the straight-line distance check and attribute consistency check one by one. If the angle exceeds the preset angle threshold, it means that the vector pair does not match; if the angle does not exceed the preset angle threshold, it is regarded as a vector pair that meets the requirements.

[0108] Step 210: Perform normal vector angle check, plane distance check, plane anti-error fitting, and attribute consistency check calculations on the surface elements.

[0109] Specifically, plane robust fitting: For the surface elements of the underlying vector or the updated vector, calculate the fitting residuals of each vector point in the underlying vector and the updated vector, calculate the root mean square (RMS), mean, and standard deviation of the fitting residuals, and mark the outlier points that deviate from the mean of the fitting residuals by two standard deviations; in the case where the RMS of the fitting residuals does not meet the preset requirements, remove the outlier points, and repeat steps 204 to 206 until the RMS of the fitting residuals meets the preset requirements, and determine the line elements of the final underlying vector or the updated vector.

[0110] Plane distance check: Calculate the distance error between each vector point in the updated vector and the surface element in the underlying vector one by one, eliminate the underlying vectors with distance errors greater than the threshold, and record the vector pairs that meet the requirements.

[0111] Attribute consistency check: Determine whether the attributes of the updated vector and the underlying vector in the vector pairs that meet the requirements determined according to the straight-line distance check are consistent. If they are inconsistent, it indicates that the vector pair does not match. If they are consistent, perform an angle consistency check on them to obtain the final vector pairs that meet the requirements.

[0112] Normal vector angle check: Calculate the angle between the normal vectors of each surface element in the updated vector and the surface element in the underlying vector in the vector pairs that meet the plane distance check and attribute consistency check one by one. In the case where the angle exceeds the preset angle threshold, it indicates that the vector pair does not match; in the case where the angle does not exceed the preset angle threshold, use it as the vector pair that meets the requirements.

[0113] In the embodiments of this specification, according to the above steps 202 to 210, the matching vector pairs in the underlying vector and the updated vector can be obtained quickly and accurately.

[0114] Step 104: Determine the pose adjustment parameters of the updated vector and the repositioning parameters of each vector point in the updated vector according to the vector matching pairs.

[0115] Specifically, the determining the pose adjustment parameters of the updated vector and the repositioning parameters of each vector point in the updated vector according to the vector matching pairs includes:

[0116] Construct a Kalman filter according to the vector matching pairs and the matching distance between the underlying vector and the updated vector in the vector matching pairs;

[0117] Calculate the pose adjustment parameters of the updated vector according to the Kalman filter;

[0118] Determine the repositioning parameters of each vector point in the updated vector according to the pose adjustment parameters and the current coordinates of the vector points in the updated vector.

[0119] Among them, Kalman filtering is an algorithm that uses the state equation of a linear system to optimally estimate the system state through system input and output observation data.

[0120] Specifically, all vector matching pairs are used as observations, the matching distance between the underlying vector and the updated vector in each vector matching pair is used as the observed quantity, and the pose adjustment parameter at the previous moment is used as the state prediction value (if it does not exist, it is set to 0), and a Kalman filter is constructed (that is, the observation equation and state equation of the Kalman filter. Among them, the observation equation and state equation of the Kalman filter are fixed filtering structures. The observation equation refers to external observations, and the state equation describes the time-domain change of the parameter to be estimated itself).

[0121] According to the constructed Kalman filter, calculate the pose adjustment parameter (such as the pose correction number) of the updated vector, and then determine the relocalization parameter (map relocalization residual) of each vector point in the updated vector according to the pose adjustment parameter and the current coordinates of the vector points in the updated vector.

[0122] In the embodiments of this specification, the pose correction number and relocalization residual of the updated vector can be calculated according to the Kalman filter. Subsequently, the coordinates of the vector points in the updated vector can be adjusted according to the pose correction number and relocalization residual of the updated vector to improve the accuracy of the coordinates of the vector points in the updated vector.

[0123] In practical applications, when the data volume of the target map is large, the target map can be segmented according to a preset segmentation rule, and then according to the segmentation result, the currency change of each segment of the map is determined. The specific implementation method is as follows:

[0124] Before constructing the Kalman filter according to the vector matching pair and the matching distance between the underlying vector and the updated vector in the vector matching pair, it further includes:

[0125] Segment the target map according to a preset segmentation rule to obtain multiple segmented maps;

[0126] Successively use each segmented map as the target segmented map, and determine the initial vector, updated vector corresponding to the target segmented map, and the vector matching pairs in the target segmented map.

[0127] Among them, the preset segmentation rule can be set according to actual applications. For example, the target map is divided according to a certain spatial range and time range to obtain multiple segmented maps.

[0128] Then, each segmented map is used as the target segmented map one by one, and the corresponding initial vector, updated vector, and vector matching pairs in the target segmented map are obtained. Subsequently, the currency change of the target segmented map can be judged according to the corresponding initial vector, updated vector, and vector matching pairs in the target segmented map.

[0129] Step 106: Adjust the updated vector according to the pose adjustment parameter and the relocalization parameter to obtain an adjusted updated vector.

[0130] Specifically, the adjusting the updated vector according to the pose adjustment parameter and the relocalization parameter to obtain an adjusted updated vector includes:

[0131] When the relocalization parameter satisfies the chi-square test, adjust the coordinates of each vector point in the updated vector according to the relocalization parameter and the pose adjustment parameter to obtain an adjusted updated vector.

[0132] Among them, the chi-square test is to statistically measure the degree of deviation between the actual observed value and the theoretically inferred value of a sample. The degree of deviation between the actual observed value and the theoretically inferred value determines the size of the chi-square value. If the chi-square value is larger, the deviation degree between the two is greater; on the contrary, the deviation between the two is smaller; if the two values are exactly equal, the chi-square value is 0, indicating that the theoretical value is completely consistent.

[0133] In specific implementation, according to the relocalization parameter of each vector point in the updated vector, perform a posteriori residual analysis on it, and judge whether the a posteriori residuals of all vector points satisfy the chi-square test. If so, adjust the coordinates of each vector point in the updated vector according to the relocalization parameter and the pose adjustment parameter to obtain an adjusted updated vector.

[0134] If the chi-square test is not passed, construct a normalized standard normal distribution residual sequence, perform variance inflation on the vectors outside 3 times the standard deviation, and reconstruct the Kalman filter and calculate the pose adjustment parameter and the relocalization parameter of the updated vector.

[0135] See Figure 3 , Figure 3 shows a flowchart for updating the vector point coordinates of the updated vector in a vector-based map processing method provided according to an embodiment of the present specification, which specifically includes the following steps.

[0136] Taking the update of the vector point coordinates of the updated vector in the target segmented map as an example, a specific introduction is made.

[0137] Step 302: Determine the base vector, updated vector of the target map, and the matching relationship between the base vector and the updated vector.

[0138] Step 304: Perform regional segmentation on the target map to obtain the target segmented map.

[0139] Specifically, for the specific implementation of performing regional segmentation on the target map to obtain the target segmented map, reference may be made to the above embodiments.

[0140] Step 306: Determine the base vector, update vector, and vector matching pairs of the target segmented map.

[0141] Step 308: Obtain the pose correction number of the previous segment.

[0142] Step 310: Construct a Kalman filter based on the base vector, update vector, and vector matching pairs of the target segmented map, and in combination with the pose correction number of the previous segment.

[0143] Specifically, use all the matching vector matching pairs in the target segmented map as observations, and use the matching distance between the base vector and the update vector in the matching pair as the observed quantity to construct the observation equation of the Kalman filter; use the pose correction number of the previous segment as the state prediction value to construct the state equation of the Kalman filter.

[0144] Step 312: Calculate the pose correction number of the update vector in the target segmented map according to the Kalman filter.

[0145] Specifically, calculating the pose correction number of the update vector in the target segmented map according to the Kalman filter can be understood as calculating the pose correction number of the update vector in the target segmented map according to the observation equation of the Kalman filter and the state equation of the Kalman filter.

[0146] Step 314: Calculate the relocalization residual of each vector point in the update vector in the target segmented map according to the pose correction number of the update vector in the target segmented map.

[0147] Step 316: Perform residual analysis on the relocalization residuals of each vector point in the update vector in the target segmented map, and determine whether the relocalization residual of each vector point satisfies the chi-square test. If so, execute Step 318; if not, execute Step 320.

[0148] Among them, the relocalization residual can be understood as the residual of the matching distance between the base vector and the update vector.

[0149] These residuals are first tested for a chi - square distribution. The chi - square distribution means that if a set is assumed to follow a normal distribution, then the sum of the squares of the elements in the set follows a chi - square distribution, and the sum of the squares of all elements should be less than a certain threshold. If the chi - square distribution fails, then the normalized standard normal distribution, i.e., (r - mean) / std, where mean is the mean and std is the standard deviation, will be calculated. Then, for the part where P{(r - mean) / std}>0.997, variance inflation is performed. The so - called variance inflation means that each observation value has a prior variance, representing the accuracy of this observation value. Now, through the test, it is found that this observation value is inaccurate, so the variance is enlarged, indicating that the accuracy of this observation value is not very reliable.

[0150] Step 318: Output the pose correction number of the updated vector in the target segmented map, as well as the relocation residuals of each vector point in the updated vector in the target segmented map, and recalculate the coordinates of the updated vector through this pose correction number and the relocation residuals.

[0151] Step 320: Construct a sequence of normalized standard normal distribution residuals, perform variance inflation on the vector points outside 3 standard deviations, and continue to execute Step 310.

[0152] In the embodiments of this specification, by using the pose correction number and the relocation residuals to adjust the vector point coordinates of the updated vector in the target segmented map, the accuracy of the vector point coordinates of the updated vector in the target map can be guaranteed subsequently, so as to more accurately determine whether the target map has undergone a currency change.

[0153] Step 108: Determine the confidence level of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pair, and the correlation relationship between the initial vector and the adjusted updated vector.

[0154] Specifically, if the target map is not segmented, the confidence level of the updated vector is directly determined according to the matching distance between the initial vector and the updated vector in the vector matching pair, and the correlation relationship between the initial vector and the adjusted updated vector.

[0155] If the target map is segmented, it is necessary to calculate the absolute accuracy, relative accuracy of the target segmented map, and the confidence level of the relative accuracy of the updated vector in the target segmented map according to its segmentation result. The specific implementation method is as follows:

[0156] The determining the confidence level of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pair, and the correlation relationship between the initial vector and the adjusted updated vector includes:

[0157] Take the matching distance between the initial vector and the updated vector in the vector matching pairs in the target segmented map as the first set;

[0158] Take the distance between the updated vector and the initial vector in the target segmented map as the second set;

[0159] Take the distance between the road geometric elements of the updated vector and the road geometric elements of the initial vector in the target segmented map as the third set;

[0160] Sample the first set, the second set, and the third set respectively, and calculate the sample mean and variance according to the sampling results;

[0161] Determine the confidence levels of the absolute accuracy and relative accuracy of the vector points in the updated vector according to the sample mean and variance of the sampling.

[0162] Specifically, before relocalization (i.e., before adjusting the coordinates of each vector point in the updated vector according to the relocalization parameters and pose adjustment parameters), take the absolute distance between all vector matching pairs in the target segmented map as the first set P1. After relocalization, take the distance between all matching vector points of all vector points of the updated vector and all vector points of the initial vector in the target segmented map as the second set P2. After relocalization, take the distance between the road geometric elements of the updated vector and the road geometric elements of the initial vector in the target segmented map as the third set P3. Randomly sample the sets P1, P2, and P3 respectively, and calculate the sample mean of each sampling. Denote the sample mean sets corresponding to each set as Q1, Q2, and Q3 respectively; calculate the sample mean and variance of Q1, Q2, and Q3, and design the interval to be detected as needed for hypothesis testing to obtain the probability value, which is the confidence level of the segmented absolute accuracy, segmented relative accuracy, and vector relative accuracy. Among them, the segmented absolute accuracy refers to the deviation degree between the obtained vector and the road elements in the real world, and the relative accuracy means the relative deviation degree. The segmented absolute accuracy, segmented relative accuracy, and vector relative accuracy correspond to Q1, Q2, and Q3 respectively, and their calculation methods are the same; and the interval to be detected can be designed according to actual needs, such as 10 cm horizontally for lane lines, 10 cm longitudinally and 30 cm horizontally and 40 cm high for traffic lights, etc.

[0163] Then the confidence levels of the absolute accuracy and relative accuracy of the vector points in the updated vector can be understood as the deviation degree between the vector points in the updated vector and the road elements in the real world. That is, the smaller the deviation degree, the less the road element in the target map has changed, and the larger the deviation degree, the more the road element in the target map has changed.

[0164] See Figure 4 , Figure 4The figure shows a flowchart for obtaining the confidence update of a vector in a vector-based map processing method provided according to an embodiment of the present specification, which specifically includes the following steps.

[0165] Step 402: Calculate the pose correction number of the updated vector based on the updated vector, the base vector, and the matching relationship between the updated vector and the base vector.

[0166] Specifically, divide the area to be detected (such as the above-mentioned target map) according to a certain spatial range and time range to obtain several segments. For each segment, read the updated vector, the base vector, and the vector matching relationship (vector matching pairs) within the segment respectively.

[0167] Use all the vector matching pairs within the segment as observations, and the matching distance between the base vector and the updated vector as the observed quantity to construct the observation equation of the Kalman filter; in the case where there is a pose correction number for the previous segment, use the pose correction number of the previous segment as the state prediction value to construct the state equation of the Kalman filter.

[0168] Calculate the pose correction number of the current segment according to the constructed Kalman filter (the observation equation of the Kalman filter and the state equation of the Kalman filter), and calculate the relocalization residual of each vector point in the updated vector according to the pose correction number of the current segment and the current coordinates of each vector point in the updated vector.

[0169] Perform posterior residual analysis based on the obtained relocalization residual. First, determine whether the posterior residuals of all vectors satisfy the chi-square test; if satisfied, output the relocalization residual, the pose correction number, and recalculate the coordinates of the updated vector.

[0170] If the chi-square test is not passed, construct a normalized standard normal distribution residual sequence, and inflate the variance of the vectors outside 3 times the standard deviation, and repeat the above steps until all data processing is completed.

[0171] Step 404: Obtain the relative distance of all-interval vectors before pose correction, the relative distance of all-interval vectors after pose correction, and the relative distance of a single vector after pose correction; perform random sampling and sampling mean calculation on the obtained relative distance of all-interval vectors before pose correction, the relative distance of all-interval vectors after pose correction, and the relative distance of a single vector after pose correction to obtain the sampling mean.

[0172] Specifically, take the absolute distance between all matching vector pairs within a certain segment before relocalization as set P1; take the distance between all matching vectors within a certain segment after relocalization as set P2; take the distance between a certain matching vector after relocalization as set P3.

[0173] Randomly sample the sets P1, P2, and P3 respectively, calculate the sample mean of each sampling, and denote the sample mean sets corresponding to each set as Q1, Q2, and Q3 respectively; calculate the sample mean and variance of Q1, Q2, and Q3.

[0174] Step 406: Calculate the confidence levels of the piecewise absolute accuracy, piecewise relative accuracy, and vector relative accuracy based on the sampling mean, and mark the changed vector points in the updated vector according to the confidence levels.

[0175] Specifically, calculate the sample mean and variance of Q1, Q2, and Q3, and design the interval to be detected as needed for hypothesis testing to obtain the probability value, which is the confidence level of the piecewise absolute accuracy, piecewise relative accuracy, and vector relative accuracy. And judge whether the vector has a currency change according to the confidence level.

[0176] Step 110: Determine whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vector.

[0177] Specifically, after determining the vector matching pairs, the pose adjustment parameters of the updated vector, and the confidence level of the updated vector, it is possible to determine whether the road elements in the target map have changed according to the above parameters.

[0178] In practical applications, there are various specific implementation methods for determining whether the road elements in the target map have changed according to the vector matching pairs, the pose adjustment parameters of the updated vector, and the confidence level of the updated vector. The following introduces two methods. The specific implementation methods are as follows:

[0179] Determining whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vector includes:

[0180] In the case where it is determined that there is no matching relationship for any vector point in the updated vector according to the vector matching pairs, determine that the target map has changed; or

[0181] In the case where it is determined that the confidence level of the updated vector does not meet the preset confidence threshold, determine that the target map has changed; or

[0182] In the case where it is determined that the pose adjustment parameters do not meet the preset accuracy threshold, determine that the target map has changed.

[0183] That is, when it is determined that there is no matching relationship for any vector point in the update vector according to the vector matching pair, it can be determined that the road elements of the target map have changed; for example, if there is a certain vector point in the initial vector but not in the update vector, it can be determined that the vector point has been deleted in the update vector; if there is a certain vector point in the update vector but not in the initial vector, it can be determined that the vector point has been added in the update vector.

[0184] Or when the confidence level of the update vector does not meet the preset confidence level threshold, it can also be determined that the target map has changed; or when it is determined that the pose adjustment parameter of the update vector does not meet the preset accuracy threshold, it is determined that the target map has changed. Among them, the preset confidence level threshold and the preset accuracy threshold can both be set according to actual applications and are not limited herein.

[0185] In addition, in another way, it is also possible to accurately determine whether the target map has changed in a progressive manner. The specific implementation method is as follows:

[0186] Determining whether the target map has changed according to the vector matching pair, the pose adjustment parameter, and / or the confidence level of the update vector includes:

[0187] When it is determined according to the vector matching pair that each vector point in the update vector has a matching relationship, determine whether the confidence level of the update vector meets the preset confidence level threshold.

[0188] If so, determine whether the pose adjustment parameter meets the preset accuracy threshold.

[0189] If not, determine that the target map has changed.

[0190] Specifically, first determine whether each vector point in the update vector has a matching relationship according to the vector matching pair. If so, determine whether the confidence level of the update vector meets the preset confidence level threshold. If not, determine that the target map has changed. If so, continue to determine whether the pose adjustment parameter meets the preset accuracy threshold. If not, determine that the target map has changed. If so, determine that the target map has not changed.

[0191] See Figure 5 , Figure 5 shows a flowchart for determining whether the target map has changed in a vector-based map processing method provided by an embodiment of this specification, which specifically includes the following steps.

[0192] Step 502: Matching relationship judgment.

[0193] Specifically, based on the vector matching pair, it is determined whether there is a matching relationship between the updated vector and the initial vector. If so, step 504 is executed; if not, step 506 is executed.

[0194] Step 504: Confidence judgment.

[0195] Specifically, the absolute precision confidence of the segment, the relative precision confidence of the segment, and the relative precision confidence of the vector are obtained, and it is determined whether the absolute precision confidence of the segment, the relative precision confidence of the segment, and the relative precision confidence of the vector meet the preset precision threshold. If so, step 508 is executed; if not, step 510 is executed.

[0196] Step 506: Vector deletion.

[0197] Specifically, vector deletion can be understood as follows: if this vector exists in the base vector but does not exist in the currently collected updated vector, it may be that this vector has disappeared in the real world, so it is marked as deleted.

[0198] Step 508: Pose correction number judgment.

[0199] Specifically, it is determined whether the pose correction number of the updated vector conforms to the prior precision provided by the trajectory. If so, step 510 is executed; if not, step 512 is executed.

[0200] Specifically, the prior precision can be understood as the positioning precision of the collection vehicle trajectory in the current segment area (target segment map). For example, if the positioning precisions of x, y, and z are 5 cm, 5 cm, and 10 cm, then it can be imagined that the pose correction number in this interval should be about the same magnitude. If the calculated pose correction number is 5 m, 5 m, and 10 m, there must be a problem.

[0201] Step 510: Vector position change.

[0202] Step 512: The vector position has not changed.

[0203] The vector-based map processing method provided by the embodiments of this specification can, based on the initial vector data and updated vector data of the target map, detect changes in the real world in a timely and accurate manner, maintain the timeliness of the data in the target map, and greatly shorten the processing cycle for the currency changes of road elements in the high-precision map by using this automated detection method, reducing costs.

[0204] Specifically, the vector-based map processing method provided in the embodiments of this specification can automatically detect changes in the real world based on vector data such as the initial vector data and updated vector data of the target map, and maintain the timeliness of map data. Compared with manual operations, it greatly improves production efficiency and reduces production costs; since vector data is composed of vector points, it can accurately describe various traffic elements such as lane lines, traffic lights, and traffic signs, and can achieve full coverage of road all-elements. In addition, compared with remote sensing images and visual pictures, vector data has more explicit three-dimensional coordinate information, and can accurately describe the location and degree of changes in the real world.

[0205] At the same time, through the above-mentioned high-precision map differential discovery method based on vector data (i.e., the vector-based map processing method), relying on vectors with accurate position information and using mathematical statistics methods, it can complete the detection of changes in the real world, with high precision, high efficiency, high reliability, and convenient implementation. And in the embodiments of this specification, technical innovations have been made in the error processing of vector matching and vector repositioning. By referring to the robust estimation method in the surveying and mapping field, it can accurately distinguish and eliminate gross errors in the observed data, and finally ensure the accuracy of the pose correction number. By combining filtering estimation, the central limit theorem, and hypothesis testing methods, etc., using the matching residuals before and after repositioning, a series of test quantities are constructed. Finally, through these test quantities, it can accurately judge the changes in the real world, and describe the reliability of the change discovery result through mathematical methods.

[0206] Corresponding to the above method embodiments, this specification also provides embodiments of a vector-based map processing device. Figure 6 The structural schematic diagram of a vector-based map processing device provided by an embodiment of this specification is shown. As Figure 6 shown, the device includes:

[0207] A vector matching module 602, configured to match the initial vector and the updated vector of the target map according to a preset matching rule, and determine vector matching pairs;

[0208] A parameter calculation module 604, configured to determine the pose adjustment parameter of the updated vector and the repositioning parameter of each vector point in the updated vector according to the vector matching pairs;

[0209] A vector adjustment module 606, configured to adjust the updated vector according to the pose adjustment parameter and the repositioning parameter to obtain an adjusted updated vector;

[0210] A confidence determination module 608, configured to determine the confidence of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pairs, and the correlation relationship between the initial vector and the adjusted updated vector;

[0211] A change determination module 610, configured to determine whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vectors.

[0212] Optionally, the vector matching module 602 is further configured to:

[0213] Obtain the initial vectors and updated vectors of the target map, and respectively determine the geometric types of the road elements corresponding to the vector points in the initial vectors and the updated vectors;

[0214] Fit the initial vectors and the updated vectors respectively through a preset fitting algorithm according to the geometric types to obtain the road geometric elements of the initial vectors and the updated vectors;

[0215] Determine vector matching pairs according to the initial vectors, the updated vectors, the road geometric elements of the initial vectors, and the road geometric elements of the updated vectors.

[0216] Optionally, the vector matching module 602 is further configured to:

[0217] Determine the fitting algorithms corresponding to the initial vectors and the updated vectors according to the geometric types;

[0218] Obtain the initial road geometric elements of the initial vectors according to the fitting algorithm corresponding to the initial vectors;

[0219] Obtain the initial road geometric elements of the updated vectors according to the fitting algorithm corresponding to the updated vectors;

[0220] Adjust the initial road geometric elements of the initial vectors and the initial road geometric elements of the updated vectors according to a preset calculation rule to obtain the road geometric elements of the initial vectors and the updated vectors.

[0221] Optionally, the vector matching module 602 is further configured to include:

[0222] Calculate the first fitting residuals of the vector points in the initial vectors according to the initial road geometric elements of the initial vectors, and calculate the root mean square, mean value, and standard deviation of the first fitting residuals according to the first fitting residuals;

[0223] Calculate the second fitting residuals of the vector points in the updated vectors according to the initial road geometric elements of the updated vectors, and calculate the root mean square, mean value, and standard deviation of the second fitting residuals according to the second fitting residuals;

[0224] Adjust the initial road geometric elements of the initial vector according to the root mean square, mean, and standard deviation of the first fitting residual to obtain the road geometric elements of the initial vector;

[0225] Adjust the initial road geometric elements of the updated vector according to the root mean square, mean, and standard deviation of the second fitting residual to obtain the road geometric elements of the updated vector.

[0226] Optionally, the vector matching module 602 is further configured to:

[0227] Calculate the distance error between the vector points of the updated vector and the road geometric elements of the initial vector, and determine the initial vector matching pairs according to the distance error;

[0228] Determine the attribute information of the vector points of the updated vector in the initial vector matching pairs, and the attribute information of the corresponding vector points of the initial vector;

[0229] When the attribute information of both is the same, calculate the included angle between the road geometric elements of the updated vector and the initial vector in the initial vector matching pairs, and determine the vector matching pairs according to the included angle.

[0230] Optionally, the parameter calculation module 604 is further configured to:

[0231] Construct a Kalman filter according to the vector matching pairs and the matching distance between the underlying vector and the updated vector in the vector matching pairs;

[0232] Calculate the pose adjustment parameters of the updated vector according to the Kalman filter;

[0233] Determine the relocalization parameters of each vector point in the updated vector according to the pose adjustment parameters and the current coordinates of the vector points in the updated vector.

[0234] Optionally, the vector adjustment module 606 is further configured to:

[0235] When the relocalization parameters satisfy the chi-square test, adjust the coordinates of each vector point in the updated vector according to the relocalization parameters and the pose adjustment parameters to obtain the adjusted updated vector.

[0236] Optionally, the device further includes:

[0237] A segmentation module, configured to:

[0238] Segment the target map according to a preset segmentation rule to obtain multiple segmented maps;

[0239] Successively take each segmented map as the target segmented map, and determine the initial vector, updated vector corresponding to the target segmented map, and the vector matching pairs in the target segmented map.

[0240] Optionally, the confidence determination module 608 is further configured to:

[0241] Take the matching distance between the initial vector and the updated vector in the vector matching pairs in the target segmented map as the first set;

[0242] Take the distance between the updated vector and the initial vector in the target segmented map as the second set;

[0243] Take the distance between the road geometric elements of the updated vector and the road geometric elements of the initial vector in the target segmented map as the third set;

[0244] Sample the first set, the second set, and the third set respectively, and calculate the sample mean and variance according to the sampling results;

[0245] Determine the confidence levels of the absolute accuracy and relative accuracy of the vector points in the updated vector according to the sample mean and variance of the sampling.

[0246] Optionally, the change determination module 610 is further configured to:

[0247] In the case where it is determined that there is no matching relationship for any vector point in the updated vector according to the vector matching pairs, determine that the target map has changed; or

[0248] In the case where it is determined that the confidence level of the updated vector does not meet the preset confidence threshold, determine that the target map has changed; or

[0249] In the case where it is determined that the pose adjustment parameter does not meet the preset accuracy threshold, determine that the target map has changed.

[0250] Optionally, the change determination module 610 is further configured to:

[0251] In the case where it is determined that there is a matching relationship for each vector point in the updated vector according to the vector matching pairs, determine whether the confidence level of the updated vector meets the preset confidence threshold,

[0252] If so, determine whether the pose adjustment parameter meets the preset accuracy threshold,

[0253] If not, determine that the target map has changed.

[0254] The vector-based map processing device provided by the embodiments of this specification can detect changes in the real world in a timely and accurate manner based on the initial vector data and updated vector data of the target map, maintain the timeliness of the data in the target map, and greatly shorten the processing cycle for the currency changes of road elements in the high-precision map by using such an automated detection method, reducing costs.

[0255] The above is a schematic solution of a vector-based map processing device according to this embodiment. It should be noted that the technical solution of the vector-based map processing device and the technical solution of the above-mentioned vector-based map processing method belong to the same concept. For the details not described in detail in the technical solution of the vector-based map processing device, reference can be made to the description of the technical solution of the above-mentioned vector-based map processing method.

[0256] Figure 7 The structural block diagram of a computing device 700 according to an embodiment of this specification is shown. The components of the computing device 700 include, but are not limited to, a memory 710 and a processor 720. The processor 720 is connected to the memory 710 through a bus 730, and a database 750 is used to store data.

[0257] The computing device 700 further includes an access device 740, which enables the computing device 700 to communicate via one or more networks 760. Examples of these networks include the Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or a combination of communication networks such as the Internet. The access device 740 may include one or more of any type of wired or wireless network interfaces (e.g., Network Interface Card (NIC)), such as an IEEE802.11 Wireless Local Area Network (WLAN) wireless interface, a Worldwide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0258] In an embodiment of this specification, the above components of the computing device 700 and Figure 7 other components not shown in Figure 7 may also be connected to each other, for example, through a bus. It should be understood that

[0259] The computing device 700 can be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.) or other types of mobile devices, or a stationary computing device such as a desktop computer or a PC. The computing device 700 can also be a mobile or stationary server.

[0260] Wherein, the processor 720 is configured to execute the following computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the above-mentioned vector-based map processing method are implemented.

[0261] The above is a schematic solution of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the above-mentioned vector-based map processing method belong to the same concept. For the details not described in detail in the technical solution of the computing device, reference can be made to the description of the technical solution of the above-mentioned vector-based map processing method.

[0262] An embodiment of this specification also provides a computer-readable storage medium, which stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the steps of the above-mentioned vector-based map processing method are implemented.

[0263] The above is a schematic solution of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the above-mentioned vector-based map processing method belong to the same concept. For the details not described in detail in the technical solution of the storage medium, reference can be made to the description of the technical solution of the above-mentioned vector-based map processing method.

[0264] An embodiment of this specification also provides a computer program, wherein when the computer program is executed on a computer, the computer is made to execute the steps of the above-mentioned vector-based map processing method.

[0265] The above is a schematic solution of a computer program according to this embodiment. It should be noted that the technical solution of this computer program and the technical solution of the above-mentioned vector-based map processing method belong to the same concept. For the details not described in detail in the technical solution of the computer program, reference can be made to the description of the technical solution of the above-mentioned vector-based map processing method.

[0266] The above describes 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 in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0267] The computer instructions include computer program code, which may be in source code form, object code form, executable file, or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, removable hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0268] It should be noted that for the foregoing method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the embodiments of this specification are not limited by the described order of actions, because according to the embodiments of this specification, certain steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments of this specification.

[0269] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0270] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not elaborate on all the details and do not limit the invention to only the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the embodiments of this specification, so that those skilled in the art can well understand and utilize this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. A vector-based map processing method, comprising: Matching the initial vector and the updated vector of the target map according to a preset matching rule to determine a vector matching pair; Determining a pose adjustment parameter of the updated vector and a relocalization parameter of each vector point in the updated vector according to the vector matching pair; wherein, the pose adjustment parameter includes a pose correction number, and the relocalization parameter includes a map relocalization residual; Adjusting the updated vector according to the pose adjustment parameter and the relocalization parameter to obtain an adjusted updated vector; Determining the confidence level of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pair and the correlation relationship between the initial vector and the adjusted updated vector; Determining whether the target map has changed according to the vector matching pair, the pose adjustment parameter and / or the confidence level of the updated vector; Wherein, the matching the initial vector and the updated vector of the target map according to a preset matching rule to determine a vector matching pair includes: Obtaining the initial vector and the updated vector of the target map, and respectively determining the geometric types of the road elements corresponding to the vector points in the initial vector and the updated vector; Fitting the initial vector and the updated vector respectively through a preset fitting algorithm according to the geometric type to obtain the road geometric elements of the initial vector and the updated vector; Determining a vector matching pair according to the initial vector, the updated vector, the road geometric element of the initial vector and the road geometric element of the updated vector.

2. The method according to claim 1, wherein the fitting the initial vector and the updated vector respectively through a preset fitting algorithm according to the geometric type to obtain the road geometric elements of the initial vector and the updated vector includes: Determining the fitting algorithms corresponding to the initial vector and the updated vector according to the geometric type; Obtaining the initial road geometric element of the initial vector according to the fitting algorithm corresponding to the initial vector; Obtaining the initial road geometric element of the updated vector according to the fitting algorithm corresponding to the updated vector; Adjusting the initial road geometric element of the initial vector and the initial road geometric element of the updated vector according to a preset calculation rule to obtain the road geometric elements of the initial vector and the updated vector.

3. The method according to claim 2, wherein the adjusting the initial road geometric element of the initial vector and the initial road geometric element of the updated vector according to a preset calculation rule to obtain the road geometric elements of the initial vector and the updated vector includes: Calculating a first fitting residual of the vector points in the initial vector according to the initial road geometric element of the initial vector, and calculating the root mean square, mean value and standard deviation of the first fitting residual according to the first fitting residual; Calculating a second fitting residual of the vector points in the updated vector according to the initial road geometric element of the updated vector, and calculating the root mean square, mean value and standard deviation of the second fitting residual according to the second fitting residual; Adjust the initial road geometric elements of the initial vector according to the root mean square, mean, and standard deviation of the first fitting residual to obtain the road geometric elements of the initial vector; Adjust the initial road geometric elements of the updated vector according to the root mean square, mean, and standard deviation of the second fitting residual to obtain the road geometric elements of the updated vector.

4. The method according to claim 1, wherein determining the vector matching pairs according to the initial vector, the updated vector, the road geometric elements of the initial vector, and the road geometric elements of the updated vector comprises: Calculate the distance error between the vector points of the updated vector and the road geometric elements of the initial vector, and determine the initial vector matching pairs according to the distance error; Determine the attribute information of the vector points of the updated vector in the initial vector matching pairs and the attribute information of the corresponding vector points of the initial vector; When the attribute information of both is the same, calculate the included angle between the road geometric elements of the updated vector and the road geometric elements of the initial vector in the initial vector matching pairs, and determine the vector matching pairs according to the included angle.

5. The method according to claim 1 or 4, wherein determining the pose adjustment parameters of the updated vector and the relocalization parameters of each vector point in the updated vector according to the vector matching pairs comprises: Construct a Kalman filter according to the vector matching pairs and the matching distance between the base vector and the updated vector in the vector matching pairs; Calculate the pose adjustment parameters of the updated vector according to the Kalman filter; Determine the relocalization parameters of each vector point in the updated vector according to the pose adjustment parameters and the current coordinates of the vector points in the updated vector.

6. The method according to claim 5, wherein adjusting the updated vector according to the pose adjustment parameters and the relocalization parameters to obtain the adjusted updated vector comprises: When the relocalization parameters satisfy the chi-square test, adjust the coordinates of each vector point in the updated vector according to the relocalization parameters and the pose adjustment parameters to obtain the adjusted updated vector.

7. The method according to claim 5, before constructing the Kalman filter according to the vector matching pairs and the matching distance between the base vector and the updated vector in the vector matching pairs, further comprises: Segment the target map according to a preset segmentation rule to obtain a plurality of segmented maps; Take each segmented map as the target segmented map in turn, and determine the initial vector, the updated vector, and the vector matching pairs in the target segmented map corresponding to the target segmented map.

8. The method according to claim 7, wherein determining the confidence level of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pairs and the correlation relationship between the initial vector and the adjusted updated vector comprises: Take the matching distance between the initial vector and the updated vector in the vector matching pairs in the target segmented map as the first set; Take the distance between the updated vector and the initial vector in the target segmented map as the second set; Take the distance between the road geometric elements of the updated vector and the road geometric elements of the initial vector in the target segmented map as the third set; Sample the first set, the second set, and the third set respectively, and calculate the sample mean and variance according to the sampling results; Determine the confidence levels of the absolute accuracy and relative accuracy of the vector points in the updated vector according to the sample mean and variance of the sampling; 9. The method according to claim 1, wherein determining whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vector includes: When it is determined according to the vector matching pairs that there is no matching relationship for any vector point in the updated vector, determine that the target map has changed; Or When it is determined that the confidence level of the updated vector does not meet the preset confidence level threshold, determine that the target map has changed; Or When it is determined that the pose adjustment parameters do not meet the preset accuracy threshold, determine that the target map has changed.

10. The method according to claim 1, wherein determining whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vector includes: When it is determined according to the vector matching pairs that there is a matching relationship for each vector point in the updated vector, determine whether the confidence level of the updated vector meets the preset confidence level threshold, If not, then determine whether the pose adjustment parameters meet the preset accuracy threshold, If not, then determine that the target map has changed.

11. A vector-based map processing device, comprising: A vector matching module configured to match the initial vector and the updated vector of the target map according to a preset matching rule to determine vector matching pairs; A parameter calculation module configured to determine the pose adjustment parameters of the updated vector and the relocation parameters of each vector point in the updated vector according to the vector matching pairs; wherein, the pose adjustment parameters include pose correction numbers, and the relocation parameters include map relocation residuals; A vector adjustment module configured to adjust the updated vector according to the pose adjustment parameters and the relocation parameters to obtain an adjusted updated vector; A confidence level determination module configured to determine the confidence level of the updated vector according to the matching distance between the initial vector and the updated vector in the vector matching pairs and the correlation relationship between the initial vector and the adjusted updated vector; A change determination module configured to determine whether the target map has changed according to the vector matching pairs, the pose adjustment parameters, and / or the confidence level of the updated vector; Wherein, matching the initial vector and the updated vector of the target map according to a preset matching rule to determine vector matching pairs includes: Obtain the initial vector and the updated vector of the target map, and respectively determine the geometric types of the road elements corresponding to the vector points in the initial vector and the updated vector; The initial vector and the updated vector are respectively fitted through a preset fitting algorithm according to the geometric type to obtain the road geometric elements of the initial vector and the updated vector; A vector matching pair is determined according to the initial vector, the updated vector, the road geometric elements of the initial vector, and the road geometric elements of the updated vector.

12. A computing device, comprising: A memory and a processor; The memory is used for storing computer-executable instructions, and the processor is used for executing the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the vector-based map processing method according to any one of claims 1-10 are implemented.

13. A computer-readable storage medium storing computer-executable instructions, which when executed by a processor implement the steps of the vector-based map processing method according to any one of claims 1-10.

Images