Lane positioning method, system, electronic device and storage medium

Abstract

The embodiment of the present invention discloses a lane positioning method, system, electronic device and storage medium. Based on the acquired vehicle lane data and vehicle trajectory data, terminal self-learning analysis is performed to obtain a first self-learning analysis result for each data selection point in the vehicle trajectory data; the first self-learning analysis result is connected in series to obtain a second self-learning analysis result of a corresponding length segment; based on the first target vehicle lane data and the vehicle trajectory data obtained by the second self-learning analysis result, first lane positioning data is obtained. Self-learning analysis is performed using historical data obtained from a selected user terminal. When the camera recognition data is inaccurate, the result obtained by the terminal self-learning analysis is used to correct the current vehicle lane data, thereby avoiding the situation where the vehicle lane is misidentified or cannot be recognized, and improving the accuracy and stability of lane-level navigation positioning.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of navigation and positioning technology, and specifically to a lane positioning method, system, electronic device and storage medium. Background Art

[0002] The current lane-level navigation and positioning technology can cause inaccurate camera recognition due to external interference with the camera (such as vehicle obstruction, unclear lane lines, backlight, weather and other factors), or lane lines being worn or obstructed due to physical reasons. This can easily lead to misidentification of lane lines or failure to identify lane lines, resulting in low lane-level navigation and positioning accuracy and poor stability.

[0003] In addition, the database data is updated slowly during the lane positioning process. The data update process includes road data collection, data compilation, data review, and data release. A data update takes at least three months. However, the lane data changes quickly. For example, if traffic rules change, the dotted lanes originally stored in the map data have actually been changed to solid lines. If the update is not timely, it will also cause lane positioning errors. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a lane positioning method, system, electronic device and storage medium to solve the problems of low accuracy and poor stability of existing lane-level navigation positioning.

[0005] To achieve the above objectives, the embodiments of the present invention mainly provide the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention provides a lane positioning method, the method comprising:

[0007] Collecting lane data and vehicle trajectory data of each vehicle, wherein the lane data includes lane type and lane color;

[0008] Perform terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data to obtain a first self-learning analysis result for each data selection point in the vehicle trajectory data;

[0009] Connecting the first self-learning analysis results of the data selection points in series to obtain a second self-learning analysis result of a line segment of a corresponding length;

[0010] Obtaining first target vehicle lane data based on the second self-learning analysis result;

[0011] First lane positioning data is obtained according to the first target vehicle lane data and the vehicle trajectory data.

[0012] In a second aspect, an embodiment of the present invention provides a lane positioning method, the method comprising:

[0013] Collecting the collection device information, lane data and vehicle trajectory data of each vehicle within a preset range, wherein the lane data includes lane type and lane color;

[0014] Performing full data self-learning analysis based on the acquired acquisition device information, vehicle line data, and vehicle trajectory data to obtain a third self-learning analysis result for each data selection point in the vehicle trajectory data;

[0015] Connecting the third self-learning analysis results of the data selection points in series to obtain a fourth self-learning analysis result of a line segment of a preset length;

[0016] Obtaining second target lane data based on the fourth self-learning analysis result;

[0017] Second lane positioning data is obtained according to the second target vehicle lane data and the vehicle trajectory data.

[0018] In a third aspect, an embodiment of the present invention provides a lane positioning method, the method comprising:

[0019] Conduct terminal self-learning analysis and full data self-learning analysis based on the vehicle's current location;

[0020] If the current position of the vehicle is learned and analyzed to obtain the first target vehicle lane data and the second target vehicle lane data;

[0021] Comparing the vehicle lane data of the current position of the vehicle with the second target vehicle lane data to see whether they are consistent;

[0022] If the two are consistent, the second lane positioning data is used as the output result.

[0023] If the two are inconsistent, the lane data of the vehicle's current position is corrected based on the fourth self-learning analysis result, and the second correction result is used as the output result.

[0024] In a fourth aspect, an embodiment of the present invention further provides a lane positioning system, the system comprising:

[0025] A first data collection module, used to collect lane data and vehicle trajectory data of each vehicle, wherein the lane data includes lane type and lane color;

[0026] A first self-learning analysis module, used to perform terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data, and obtain a first self-learning analysis result of each data selection point in the vehicle trajectory data;

[0027] A second self-learning analysis module, used to connect the first self-learning analysis results of a preset number of data selection points in series to obtain a second self-learning analysis result of a line segment of a preset length;

[0028] A first determination module, configured to obtain first target vehicle lane data based on the second self-learning analysis result;

[0029] The first lane positioning module is used to obtain first lane positioning data according to the first target vehicle lane data and the vehicle trajectory data.

[0030] In a fifth aspect, an embodiment of the present invention further provides a lane positioning system, the system comprising:

[0031] A second data collection module is used to collect the collection device information, vehicle lane data and vehicle trajectory data of each vehicle within a preset range, wherein the vehicle lane data includes the vehicle lane type and the vehicle lane color;

[0032] A third self-learning analysis module is used to perform full data self-learning analysis based on the acquired acquisition device information, vehicle line data and vehicle trajectory data, and obtain a third self-learning analysis result of each data selection point in the vehicle trajectory data;

[0033] A fourth self-learning analysis module, used to connect the third self-learning analysis results of a preset number of data selection points in series to obtain a fourth self-learning analysis result of a line segment of a preset length;

[0034] A second determination module, configured to obtain second target lane data based on the fourth self-learning analysis result;

[0035] The second lane positioning module is used to obtain second lane positioning data according to the second target vehicle lane data and the vehicle trajectory data.

[0036] In a sixth aspect, an embodiment of the present invention further provides an electronic device, comprising: at least one processor and at least one memory;

[0037] The memory is used to store one or more program instructions;

[0038] The processor is used to run one or more program instructions to execute a lane positioning method as described in the first aspect to the third aspect.

[0039] In a seventh aspect, an embodiment of the present invention further provides a computer-readable storage medium comprising one or more program instructions, wherein the one or more program instructions are used to execute a lane positioning method as described in the first aspect to the third aspect.

[0040] The technical solution provided by the embodiment of the present invention has at least the following advantages:

[0041] The lane positioning method, system, electronic device and storage medium provided by the embodiment of the present invention, based on the vehicle lane positioning technology, uses network big data to achieve fast intelligent self-learning, and can fully learn, so as to intelligently provide vehicle lane positioning, and improve the accuracy of vehicle lane positioning. Specifically, the terminal self-learning analysis is performed using the historical data obtained from the selected user terminal. When the camera recognition data is inaccurate, the result obtained by the self-learning analysis is used to correct the current vehicle lane data, thereby avoiding the situation where the vehicle lane is misidentified or cannot be recognized, and improving the accuracy and stability of lane-level navigation positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a lane positioning method disclosed in the first embodiment of the present invention Figure 1 .

[0043] Figure 2 A schematic diagram of a lane positioning method disclosed in the first embodiment of the present invention Figure 2 .

[0044] Figure 3 It is a flowchart of performing terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data in the first embodiment of the present invention.

[0045] Figure 4 This is a schematic diagram of a process for obtaining the first target vehicle lane data based on the second self-learning analysis result in the first embodiment of the present invention.

[0046] Figure 5 A schematic diagram of lane data of a certain road section stored in a user terminal disclosed in an embodiment of the present invention.

[0047] Figure 6 This is a schematic diagram of a lane on a road section after the lane data has changed according to an embodiment of the present invention.

[0048] Figure 7 It is a flow chart of a lane positioning method disclosed in the second embodiment of the present invention.

[0049] Figure 8 It is a flowchart of performing terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data in the second embodiment of the present invention.

[0050] Fig. 9 It is a flowchart of a lane positioning method disclosed in the third embodiment of the present invention.

[0051] Fig.10 It is a structural schematic diagram of a lane positioning system disclosed in the fourth embodiment of the present invention.

[0052] Fig.11It is a structural schematic diagram of a lane positioning system disclosed in the fifth embodiment of the present invention.

[0053] Fig.12 It is a schematic structural diagram of an electronic device disclosed in the fifth embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following is a description of the implementation of the present invention by means of specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0055] In the description of the present invention, it should be understood that the term "plurality" means at least one, may be two, or may be more than two. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0056] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood in specific circumstances.

[0057] The current lane-level navigation and positioning technology simply combines the accuracy of map data with camera-recognized vehicle lane data to perform lane positioning, which often results in vehicle lane positioning errors.

[0058] First aspect

[0059] The embodiment of the present invention provides a lane positioning method, such as Figure 1 As shown, the method comprises the following steps:

[0060] S11: Collecting lane data and vehicle trajectory data of each vehicle, wherein the lane data includes lane type and lane color;

[0061] S12: performing terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data to obtain a first self-learning analysis result for each data selection point in the vehicle trajectory data;

[0062] S13: connecting the first self-learning analysis results of the data selection points in series to obtain a second self-learning analysis result of a line segment of a corresponding length;

[0063] S14: obtaining first target lane data based on the second self-learning analysis result;

[0064] S15: Obtaining first lane positioning data according to the first target vehicle lane data and the vehicle trajectory data.

[0065] In an embodiment of the present invention, terminal self-learning analysis is performed based on the data collected from each vehicle to obtain the learning result of each data selection point in each vehicle trajectory, whether the learning of the point is successful or failed, and the attributes of the point. The first self-learning result obtained by learning is connected in series to obtain the second self-learning analysis result of the line segment of the corresponding length, whether the learning of the line segment is successful or failed, and the attributes of the line segment. Finally, the vehicle line attributes of each line segment are obtained, and then the lane positioning data is obtained in combination with the vehicle trajectory.

[0066] Furthermore, in this embodiment, if Figure 2 As shown, the method also includes the following processing steps:

[0067] S21: Obtaining the lane data of the current position of the vehicle;

[0068] S22: comparing the vehicle lane data of the current position of the vehicle with the first target vehicle lane data to see whether they are consistent;

[0069] S23: If the two are consistent, the first lane positioning data is used as an output result;

[0070] S24: If the two are inconsistent, the lane data of the current position of the vehicle is corrected based on the second self-learning analysis result, and the first correction result is used as the output result.

[0071] The above-mentioned embodiment of the present invention uses network big data on the basis of the vehicle lane positioning technology to achieve fast intelligent self-learning, and can fully learn, so as to intelligently provide vehicle lane positioning and improve the accuracy of vehicle lane positioning. Specifically, the terminal self-learning analysis is performed using the historical data obtained from the selected user terminal. When the camera recognition data is inaccurate, the result obtained by the self-learning analysis is used to correct the current vehicle lane data, thereby avoiding the situation where the vehicle lane is misidentified or cannot be recognized, and improving the lane-level navigation positioning accuracy and stability.

[0072] Furthermore, in this embodiment, if Figure 3 As shown, before the terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data is performed, the following steps S31-S33 need to be performed, and the first self-learning analysis result of each data selection point in the vehicle trajectory is further obtained, and the following steps S34-S37 are continued:

[0073] S31: Counting the first occurrence times of each vehicle lane data of each data selection point in the vehicle trajectory data;

[0074] S32: Determine whether the first occurrence number of each lane data satisfies the first threshold value, if not, jump to step S33, if yes, jump to step S34;

[0075] S33: saving each lane data corresponding to each data selection point as a first intermediate result;

[0076] S34: Counting the first weight of each lane data at each data selection point, where the first weight is the occurrence probability of each lane data;

[0077] S35: Determine whether the first weight satisfies a second threshold value, if yes, jump to step S36, if no, jump to step S37;

[0078] S36: The data selection point is learned successfully, and the lane data corresponding to the data selection point is used as the first self-learning analysis result;

[0079] S37: The data selection point learning failed.

[0080] Specifically, the following examples are listed for illustration. Taking the vehicle lane type as an example, the vehicle lane types of each point include solid line, dotted line, dotted line, solid dotted line, double solid line, etc. In the vehicle trajectory data, a point can be marked every meter to obtain the data selection point. After a period of accumulation, the data selection point is passed through many times, and each time it is passed, a point is marked, and the first occurrence number and the first weight of each vehicle lane type marked each time are counted. For example, the first occurrence number of each vehicle lane type of the point A is: solid line (M = 100 times), dotted line (N = 9 times), dotted line (L = 1 time), solid dotted line (K = 1 time), double solid line (I = 0 times), that is, it is passed 111 times, of which 100 times are marked as solid line, 9 times are marked as dotted line, 1 time is marked as dotted line, and 1 time is marked as solid dotted line. Then, according to the above marking records, the learning result of the point is calculated. If the threshold of the number of times a certain type of vehicle line reaches 50 times (custom experience value), and the total proportion reaches 90%, the above is the first preset condition. For example, if point A: (M / (M+N+K+L+I))≥90%, then it is considered that the learning of point A is successful, and the learning result of point A is a solid line. The learning result of each point is obtained, and finally the learning results of each point are strung together to obtain the learning result of the vehicle line. If the first number of occurrences of each vehicle line data meets the first threshold, if the threshold of the number of times a certain type of vehicle line is less than 50 times (custom experience value), then the vehicle line data corresponding to each data selection point will be saved as the first intermediate result, for the next time the terminal self-learning analysis will be continued based on the first intermediate self-learning analysis result.

[0081] Furthermore, in this embodiment, the first self-learning analysis results of the data selection points are connected in series to obtain the second self-learning analysis results of the corresponding length segments. Specifically, the data selection points that have the same and continuous first self-learning analysis results are connected in series to obtain the learning results of each continuous sub-segment as the second self-learning analysis results.

[0082] Further, in this embodiment, the first target lane data is obtained based on the second self-learning analysis result, such as Figure 4 As shown, the specific steps are as follows:

[0083] S41: Determine a learning result judgment threshold according to the road type;

[0084] S42: Determine whether the number of data selection points included in the continuous sub-segment is greater than or equal to the learning result judgment threshold, if so, jump to step S43, if not, jump to step S46;

[0085] S43: confirm whether the learning result of the continuous sub-segment is successful, if yes, jump to step S44, if no, jump to step S45;

[0086] S44: taking the lane data of the continuous sub-segment as the first target lane data;

[0087] S45: acquiring the vehicle lane data stored in the line segment in the map database as the first target vehicle lane data;

[0088] S46: Determine whether the number of data selection points contained in the continuous sub-segments adjacent to and before the continuous sub-segment is greater than or equal to the learning result judgment threshold, if so, jump to step S47, if not, jump to step S48;

[0089] S47: The first half and the second half of the continuous sub-segment respectively use the lane data of the adjacent continuous sub-segment as the first target lane data;

[0090] S48: Acquire the vehicle lane data stored in the line segment in the map database as the first target vehicle lane data.

[0091] In the above steps, specifically, if the number of data selection points included in the continuous sub-segment adjacent to the continuous sub-segment before and / or after is less than the learning result judgment threshold, the vehicle lane data stored in the segment in the map database is obtained as the first target vehicle lane data. There are three situations, that is, if any one of the adjacent items before or after the continuous sub-segment or both of the adjacent items before and after is less than the learning result judgment threshold, the vehicle lane data stored in the segment in the map database is obtained as the first target vehicle lane data.

[0092] Furthermore, in actual use, the learning result judgment threshold may be different depending on the road type. For example, if the interval between two solid lines of a dashed line on a highway is 9 meters, the learning result judgment threshold is determined to be 10 points. If it is a dashed line on an ordinary road, the interval between the two solid lines is 6 meters, so the learning result judgment threshold is determined to be 7 points. The learning result judgment threshold is generally greater than the length of the interval.

[0093] That is, when conducting specific learning analysis, on the highway, because the interval between two solid lines is 9 meters, the learning result judgment threshold determined here is 10 points (one point is marked at an interval of 1 meter), and the continuous learning success unit is at least a group of 10 points, then:

[0094] When the learning of 10 or less consecutive data selection points fails, the learning result of the continuous sub-segment is divided equally according to the learning results of the two continuous sub-segments before and after it. For example, if the learning of points 0-20 is successful, points 21-25 are unsuccessful, and points 26-46 are successful, and if the line type of points 0-20 and 26-46 are both solid lines, then the interval 0-46 is a solid line; if the line type of points 0-20 is a solid line, the line type of points 26-46 is a dotted line, and there is no learning result for points 21-25, it is judged that this section is a transition stage. After the learning is completed, the two intervals before and after 21-25 are divided equally, then 0-23 is a solid line, and 24-46 is a dotted line.

[0095] When more than 10 consecutive data selection points fail to learn, the learning results of the continuous sub-line segment use the data stored in the database data (DB). For example, if 0-3 points succeed, 4th point fails, 6-10 points succeed, 11th point fails, and 12-23 succeed, then 0-11 points have no learning results, 12-23 have learning results, and the 0-11 points that have not been successfully learned use the data stored in the database data (DB), and the 12-23 successful segment still uses the learning result of the car line type. For another example, 0-20 points are successfully learned, 21-32 points are unsuccessful, and 33-53 points are successfully learned. Because 21-32 has not been successfully learned, more than 10 points have not been successfully learned at this time. If the 21-32 point segment has not been successfully learned, the data stored in the database data (DB) is used, that is, the 0-20 point segment is the learning result of the car line type, the 21-32 point segment uses the database data (DB) The car line type, and the 33-53 point segment still uses the learning result of the car line type.

[0096] Furthermore, the method further comprises:

[0097] Acquire the collection equipment information of each vehicle; associate and store the collection equipment information of each vehicle with the first target vehicle lane data of the vehicle obtained based on the second self-learning analysis result. Figure 5As shown, for the fourth car line, it is shown as a dotted line in the figure. Different acquisition devices (cameras) have different recognition effects. If the camera can easily recognize the dotted line as a solid line, after big data comparison, it is found that 90% of the users of this camera (the number of users needs to be more than 50) recognize the fourth car line at this location as a solid line. Then the camera hardware model and the car line type of the fourth car line are saved as a solid line in the production cloud data. The next time such users recognize the car line type of the fourth car line at this location as a solid line, although it does not match the stored dotted line, it can match the cloud data. It is also considered a successful match. The car line positioning error will not be caused by the wrong matching of the car line type of the car line.

[0098] Figure 5 The schematic diagram of the lane data of a certain road section stored by a user terminal disclosed in one embodiment of the present invention. The figure takes the lane type as an example for explanation. The lane information of the road section stored by the user terminal is that there are four lanes and five lanes from left to right. The four lanes are the first to fourth lanes, and the five lanes are the first to fifth lanes. The database data of the lane type corresponding to the first to fifth lanes are stored as solid line, dotted line, dashed line, dotted line, solid line. The arrows in the figure represent vehicles, that is, when the vehicle is traveling in the second lane, if more than 80% of the user terminals recognize the lane types of the first to fourth lanes as solid line, dotted line, dashed line, solid line. At this time, the lane types (solid line, dotted line, dashed line) corresponding to the first to third lanes can match the stored lane types. The embodiment of the present invention utilizes the above-mentioned terminal self-learning analysis, and accumulates the number of big data samples. Once there are more than 200 users, the lane types corresponding to the first to third lanes of the above-mentioned road section are all identified as solid lines, dotted lines, and dashed lines. Then, the lane types corresponding to the first to third lanes are recorded in the cloud data as the self-learning analysis results of historical data. When an individual user's vehicle is driving in the second lane from the left, if the first lane is mistakenly identified as a dotted line due to vehicle obstruction, rain obstruction, temporary obstacle interference, etc., then by comparing the data learned in the cloud of the road section, the information of the first lane misidentification is corrected. Thereby preventing lane positioning errors caused by camera misidentification. On the basis of lane positioning technology, network big data is used to achieve fast and intelligent self-learning, and it can fully learn, so as to intelligently give lane positioning and improve the accuracy of lane positioning.

[0099] And as Figure 6 As shown, similarly, the lane information of the road section stored in the user terminal is that there are four lanes and five lanes from left to right, the four lanes are the first to fourth lanes, and the five lanes are the first to fifth lanes. Figure 5, assuming that the third lane has changed from a dotted line to a solid line due to changes in traffic regulations, but due to the slow update of database data collection, the lane types corresponding to the first to fifth lanes in the database data (DB) used by the user are still the same as shown in the figure above Figure 5 As shown in (solid line, dotted line, dashed line, dotted line, solid line), after the user changes the line type, the probability of positioning error at the location will be greatly increased. Therefore, by collecting the line type of the third line, after learning and analysis, the line type of the third line is changed from a stable dotted line to a stable solid line. The adjusted line type of the third line is updated to the cloud data. The next time the terminal recognizes it as a solid line, it can be matched through the cloud data, thereby improving the accuracy of positioning.

[0100] Second aspect

[0101] The embodiment of the present invention provides a lane positioning method, such as Figure 7 As shown, the method comprises the following steps:

[0102] S51: Collecting the collection device information, vehicle lane data and vehicle trajectory data of each vehicle within a preset range, wherein the vehicle lane data includes the vehicle lane type and vehicle lane color;

[0103] S52: performing full data self-learning analysis based on the acquired acquisition device information, vehicle line data and vehicle trajectory data to obtain a third self-learning analysis result for each data selection point in the vehicle trajectory data;

[0104] S53: connecting the third self-learning analysis results of the data selection points in series to obtain a fourth self-learning analysis result of a line segment of a preset length;

[0105] S54: obtaining second target lane data based on the fourth self-learning analysis result;

[0106] S55: Obtain second lane positioning data according to the second target vehicle lane data and the vehicle trajectory data.

[0107] In an embodiment of the present invention, terminal self-learning analysis is performed based on the data collected from each vehicle to obtain the learning result of each data selection point in each vehicle trajectory, whether the learning of the point is successful or failed, and the attributes of the point. The first self-learning result obtained by learning is connected in series to obtain the second self-learning analysis result of the line segment of the corresponding length, whether the learning of the line segment is successful or failed, and the attributes of the line segment. Finally, the vehicle line attributes of each line segment are obtained, and then the lane positioning data is obtained in combination with the vehicle trajectory.

[0108] Furthermore, in this embodiment, the method further includes:

[0109] The lane data of the vehicle's current position is obtained; the lane data of the vehicle's current position is compared with the second target lane data to see if they are consistent; if the two are consistent, the second lane positioning data is used as the output result; if the two are inconsistent, the lane data of the vehicle's current position is corrected based on the fourth self-learning analysis result, and the second correction result is used as the output result. Specifically, the processing process of this part is the same as the processing process of correcting the second self-learning analysis result in Example 1, and will not be repeated here.

[0110] Furthermore, in this embodiment, if Figure 8 As shown, before the full data self-learning analysis based on the acquired acquisition device information, vehicle line data and vehicle trajectory data is performed, the following steps S61-S63 need to be performed, and the third self-learning analysis result of each data selection point in the vehicle trajectory data is further obtained, and the following steps S64-S637 are continued:

[0111] S61: Counting the second occurrence times of each vehicle lane data of each data selection point in the vehicle trajectory data;

[0112] S62: Determine whether the second occurrence number of each lane data satisfies the third threshold value, if not, jump to step S63, if yes, jump to step S64;

[0113] S63: saving the lane data of each data selection point as a second intermediate result;

[0114] S64: Counting the second weights of each lane data at each data selection point, where the second weights are the accumulated weights of each lane data;

[0115] S65: Determine whether the second weight satisfies a fourth threshold, if yes, jump to step S66, if no, jump to step S67;

[0116] S66: The data selection point is learned successfully, and the lane data corresponding to the data selection point is used as the third self-learning analysis result;

[0117] S67: The data selection point learning failed.

[0118] Furthermore, in the present embodiment, the second weight of each vehicle lane data of each data selection point is specifically as follows: based on each vehicle lane data of each data selection point in the vehicle trajectory data, the third number of occurrences of each type of acquisition equipment is counted; and the third weight corresponding to each type of acquisition equipment is obtained; and then the weight of each type of acquisition equipment is obtained according to the product of the third number of occurrences of each type of acquisition equipment and the corresponding third weight; finally, the sum of the weights of each type of acquisition equipment is calculated to obtain the second weight.

[0119] Specifically, each lane has corresponding sampled data, and the lane data identified by the multi-function camera (MFK) is always the same as the lane data stored in the database data (DB). A point is made every meter. The method is similar to the terminal self-learning analysis, but the camera hardware type weight is added. That is, taking the lane type as an example, the lane type of each point has a solid line, a dotted line, a dotted line, a solid dotted line, a double solid line, etc., wherein the second weight is a different parameter assignment given according to the quality of the hardware, which is a custom experience value. Through the above embodiment, the cumulative weights of each lane type are obtained by calculation; for example, the cumulative weights of each lane type at a certain point are: solid line: M = 100 times × hardware A (0.7) + 120 times × hardware B (0.5), dotted line: N = 9 times × hardware A (0.7) + 7 times × hardware C (0.3), dotted line: L = 20 times × hardware C (0.3), solid dotted line: K = 5 times × hardware D (0.1), double solid line: I = 0 times. The cumulative weight of each vehicle line type reaches 200 (this item is a custom experience value), and the cumulative weight accounts for 90% of the overall weight, that is, (M / (M+N+L+K+I)))≥90%. The above is the preset condition. At this time, the corresponding vehicle line data is used as the third self-learning analysis result of the data selection point.

[0120] Further, in this embodiment, the third self-learning analysis result of each data selection point is connected in series to obtain the fourth self-learning analysis result of the corresponding length segment, specifically: the continuous data selection points with the same third self-learning analysis result are connected in series to obtain the learning results of each continuous sub-segment as the fourth self-learning analysis result. Specifically, the processing process of this part is the same as the processing process of connecting the first self-learning analysis result in series in the first embodiment, and will not be repeated here.

[0121] Further, in this embodiment, the second target vehicle lane data is obtained based on the fourth self-learning analysis result, specifically: determining a learning result judgment threshold according to the road type;

[0122] When the number of data selection points contained in a continuous sub-segment is greater than or equal to the learning result judgment threshold; and the learning result of the continuous sub-segment is successful, the vehicle lane data of the continuous sub-segment is used as the second target vehicle lane data; or if the learning result of the continuous sub-segment is a failure, the vehicle lane data stored in the segment in the map database is obtained as the second target vehicle lane data.

[0123] The obtaining of the second target lane data based on the fourth self-learning analysis result also includes:

[0124] When the number of data selection points contained in a continuous sub-segment is less than the learning result judgment threshold; and the number of data selection points contained in the continuous sub-segments adjacent to the front and back of the continuous sub-segment is greater than or equal to the learning result judgment threshold, the front half and the second half of the continuous sub-segment respectively adopt the vehicle line data of its adjacent continuous sub-segment as the second target vehicle line data; when the number of data selection points contained in the continuous sub-segment adjacent to the front and / or back of the continuous sub-segment is less than the learning result judgment threshold, the vehicle line data stored in the segment in the map database is obtained as the second target vehicle line data.

[0125] In an embodiment of the present invention, when there is a problem with the camera, the camera recognition error can be corrected by the data learned by other cameras. For example, for the type of the same lane, the database data is stored as a solid line. For example, there are 200 vehicle acquisition devices (cameras), 195 of which recognize solid lines, and 5 other lines. Then the cameras of these 5 cameras are most likely not verified and learned, or there is a problem with the installation angle. If there is a problem with the database data (DB), the problem can be corrected by the data learned by the overall camera. For example, for the type of the same lane, the database data (DB) is stored as a dotted line. There are 200 vehicle acquisition devices (cameras), 195 of which recognize solid lines, and 5 are dotted lines. Then the type of the lane of this lane is most likely a problem with the stored data, not a problem with the camera. When there is no problem with the vehicle acquisition device (camera), there is no problem with the database data (DB), but the lane is physically worn. For example, for the same lane, the database data (DB) stores a dotted line, but because the lane is worn, the camera cannot recognize it. For example, there are 200 vehicle collection devices (cameras), 195 of which fail to be recognized, and 5 are dotted lines. It is considered that there is a problem with the vehicle line itself, which can also be solved by analyzing the data rules with big data in the above embodiment.

[0126] The third aspect

[0127] The embodiment of the present invention provides a lane positioning method, such as Fig. 9 As shown, the method comprises the following steps:

[0128] S71: Perform terminal self-learning analysis and full data self-learning analysis based on the current position of the vehicle;

[0129] S72: if the current position of the vehicle is analyzed and the first target vehicle lane data and the second target vehicle lane data are obtained;

[0130] S73: comparing the vehicle lane data of the current position of the vehicle with the second target vehicle lane data to see whether they are consistent, if so, jumping to step S74, if not, jumping to step S75;

[0131] S74: taking the second lane positioning data as an output result;

[0132] S75: Correcting the vehicle lane data of the vehicle's current position based on the fourth self-learning analysis result, and using a second correction result as an output result.

[0133] In the embodiment of the present invention, when the results of cloud-based learning (full data self-learning analysis) are sent to terminals of the same hardware type, the terminals continue to learn directly based on the results of cloud-based learning (terminal self-learning analysis), thereby solving the problem that the vehicle has not been learned when it is just produced and the vehicle line positioning is slightly poor.

[0134] In addition, after the learning and analysis is completed, you can also test and verify whether the learning results are usable. This mainly involves taking the trajectory recently uploaded by the customer and adding the learned vehicle line data. Through the learning and analysis results, you can confirm that there is no abnormal lane jumping action in the lane positioning, and then test the learning and analysis results.

[0135] For the road sections complained by users, we analyze and compare the data. For example, most vehicles with other hardware types using the same DB do not jump the lane, which proves that there is no problem with the lane at that location, but that the vehicle’s camera may have a problem recognizing the road conditions at that location. Then we statistically analyze the results that the camera often recognizes, use only the user’s camera data for learning, and make the data here into a camera recognition type, so that the designated vehicle can accurately locate the lane at that location. If other hardware types of vehicles use the same DB and there are many cases of jumping the lane, it may be that the lane has changed, or there is physical wear and tear, or there are obstacles blocking it, etc., so we refer to the data of other vehicles to learn accurate lane data. Alternatively, trajectory data near the problem points complained by users may be extracted for manual analysis. Based on the analysis results, the reason why the lane is often mispositioned may be determined. After checking the camera and vehicle lane data, if it is due to hardware or vehicle lane data problems, the vehicle lane data collected manually (including collection by staff, upload by rewarded users, and actual vehicle lane data obtained from map vendors) may be compared to confirm that it is a data problem. The vehicle lane data for the area may be manually generated. If it is determined to be a camera problem, the camera characteristics may be manually analyzed, and data may be generated based on the camera special effects. If it is a lane positioning algorithm problem, it may be transferred to the lane positioning algorithm for resolution, and the correction results may be finally generated and sent to the terminal of the complaining user.

[0136] Furthermore, based on the embodiments listed in the above three aspects, the lane data further includes the number of lanes, lane width, lane width, lane curvature and curvature change rate, the positions of stop lines, zebra crossings, and road arrows. Based on the acquired attributes of the stop lines, zebra crossings, road arrows, and the directions of the road arrows for straight lines, left turns, U-turns, straight lines plus U-turns, and straight lines plus left turns, different learning methods are used for processing each attribute. Specifically,

[0137] The lane data includes the number of lanes, lane width, lane width, lane curvature and curvature change rate, and the fifth self-learning analysis result is obtained by learning through the Kalman filter algorithm based on the number of lanes, lane width, lane width, lane curvature and curvature change rate of each data selection point obtained; or

[0138] The lane data includes the positions of the stop line, zebra crossing, and road arrows, and based on the acquired longitude and latitude values ​​of the positions of the stop line, zebra crossing, and road arrows, a sixth self-learning analysis result is obtained by learning through an average value method; or

[0139] The lane data includes road arrow directions for a straight line, left turn, U-turn, straight line plus U-turn, and straight line plus left turn. The number of each type of road arrow direction for each data selection point obtained is accumulated respectively, and the result with the largest number of times is counted as the seventh self-learning analysis result.

[0140] The fourth aspect

[0141] The embodiment of the present invention also provides a lane positioning system, such as Fig.10 As shown, the system comprises:

[0142] A first data collection module 80, for collecting lane data and vehicle trajectory data of each vehicle, wherein the lane data includes lane type and lane color;

[0143] A first self-learning analysis module 81, configured to perform terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data, and obtain a first self-learning analysis result for each data selection point in the vehicle trajectory data;

[0144] A second self-learning analysis module 82, configured to connect the first self-learning analysis results of a preset number of data selection points in series to obtain a second self-learning analysis result of a line segment of a preset length;

[0145] A first determination module 83, configured to obtain first target lane data based on the second self-learning analysis result;

[0146] The first lane positioning module 84 is used to obtain first lane positioning data according to the first target vehicle lane data and the vehicle trajectory data.

[0147] Corresponding to the lane positioning method disclosed above, an embodiment of the present invention further discloses a lane positioning system. It should be noted that the specific implementation of the lane positioning system of this embodiment is similar to the specific implementation of the lane positioning method of the embodiment of the present invention. For details, please refer to the description of the lane positioning method part. In order to reduce redundancy, it will not be repeated.

[0148] Fifth Aspect

[0149] The embodiment of the present invention also provides a lane positioning system, such as Fig.11 As shown, the system comprises:

[0150] A second data collection module 90, for collecting collection device information, vehicle lane data and vehicle trajectory data of each vehicle within a preset range, wherein the vehicle lane data includes vehicle lane type and vehicle lane color;

[0151] A third self-learning analysis module 91 is used to perform full data self-learning analysis based on the acquired acquisition device information, vehicle line data and vehicle trajectory data, and obtain a third self-learning analysis result for each data selection point in the vehicle trajectory data;

[0152] A fourth self-learning analysis module 92, configured to connect the third self-learning analysis results of a preset number of data selection points in series to obtain a fourth self-learning analysis result of a line segment of a preset length;

[0153] A second determination module 93, configured to obtain second target lane data based on the fourth self-learning analysis result;

[0154] The second lane positioning module 94 is used to obtain second lane positioning data according to the second target vehicle lane data and the vehicle trajectory data.

[0155] Corresponding to the lane positioning method disclosed above, an embodiment of the present invention further discloses a lane positioning system. It should be noted that the specific implementation of the lane positioning system of this embodiment is similar to the specific implementation of the lane positioning method of the embodiment of the present invention. For details, please refer to the description of the lane positioning method part. In order to reduce redundancy, it will not be repeated.

[0156] The sixth aspect

[0157] An embodiment of the present invention further provides an electronic device, comprising: at least one processor and at least one memory;

[0158] like Fig.12 As shown, the electronic device includes a processor 100, a memory 110, an input device 120, and an output device 130; the number of processors 100 in the electronic device can be one or more. Fig.12 A processor 100 is taken as an example; the processor 100, the memory 110, the input device 120 and the output device 130 in the electronic device can be connected via a bus or other means. Fig.12 The example of connecting through bus is taken in the following.

[0159] The memory 110 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the dynamic library processing method in the embodiment of the present invention. The processor 100 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 110, that is, implements the above-mentioned dynamic library processing method.

[0160] The memory 110 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 110 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 110 may further include a memory remotely arranged relative to the processor 100, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0161] Seventh Aspect

[0162] An embodiment of the present invention further provides a computer-readable storage medium, comprising one or more program instructions, wherein the one or more program instructions are used to execute a lane positioning method as described in the first aspect to the third aspect.

[0163] Of course, the computer executable instructions of a storage medium containing computer executable instructions provided by an embodiment of the present invention are not limited to the operations of the method described above, and can also execute related operations in the dynamic library processing method provided by any embodiment of the present invention.

[0164] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0165] Those skilled in the art will appreciate that in one or more of the above examples, the functions described in the present invention can be implemented using a combination of hardware and software. When software is used, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any media that facilitates the transmission of computer programs from one place to another. Storage media can be any available media that can be accessed by general or special-purpose computers.

[0166] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A lane positioning method, characterized in that: The method comprises: Collecting lane data and vehicle trajectory data of each vehicle, wherein the lane data includes lane type and lane color; Performing vehicle terminal self-learning analysis based on the acquired vehicle line data and vehicle trajectory data to obtain a first self-learning analysis result for each data selection point in the vehicle trajectory data; Connecting the first self-learning analysis results of the data selection points in series to obtain a second self-learning analysis result of a line segment of a corresponding length; Obtaining first target vehicle lane data based on the second self-learning analysis result; Obtaining first lane positioning data according to the first target vehicle lane data and current vehicle trajectory data; Get the lane data of the vehicle's current position; Comparing the vehicle lane data of the current position of the vehicle with the first target vehicle lane data to see whether they are consistent; If the two are consistent, the first lane positioning data is used as the output result; if the two are inconsistent, the lane data of the current position of the vehicle is corrected based on the second self-learning analysis result, and the first correction result is used as the output result; The vehicle terminal self-learning analysis is performed based on the acquired vehicle line data and vehicle trajectory data to obtain a first self-learning analysis result for each data selection point in the vehicle trajectory, specifically: Count the first occurrence times of each vehicle lane data of each data selection point in the vehicle trajectory data; When it is determined that the first occurrence number of each lane data satisfies the first threshold, a first weight of each lane data at each data selection point is counted, where the first weight is the occurrence probability of each lane data; Determining whether the first weight satisfies a second threshold; If yes, the data selection point is learned successfully, and the lane data corresponding to the data selection point is used as the first self-learning analysis result; If not, the learning of the data selection point fails.

2. A lane positioning method according to claim 1, characterized in that: Before performing vehicle terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data, the method further includes: Determining whether a first occurrence number of each lane data satisfies a first threshold; If not, each lane data corresponding to each data selection point is saved as the first intermediate result.

3. A lane positioning method according to claim 1, characterized in that: The first self-learning analysis results of the data selection points are connected in series to obtain the second self-learning analysis results of the corresponding length segments, specifically: The data selection points having the same and continuous first self-learning analysis results are connected in series to obtain the learning results of each continuous sub-segment as the second self-learning analysis result.

4. A lane positioning method according to claim 3, characterized in that: The first target lane data is obtained based on the second self-learning analysis result, specifically: Determine the learning result judgment threshold according to the road type; When the number of data selection points contained in the continuous sub-segments is greater than or equal to the learning result judgment threshold; If the learning result of the continuous sub-segment is successful, the vehicle lane data of the continuous sub-segment is used as the first target vehicle lane data; Or if the learning result of the continuous sub-segment is failure, the vehicle lane data stored in the segment in the map database is obtained as the first target vehicle lane data.

5. A lane positioning method according to claim 4, characterized in that: The step of obtaining the first target lane data based on the second self-learning analysis result further includes: When the number of data selection points contained in the continuous sub-segments is less than the learning result judgment threshold; and The number of data selection points contained in the continuous sub-segments adjacent to and before the continuous sub-segment is greater than or equal to the learning result judgment threshold, and the first half and the second half of the continuous sub-segment respectively use the vehicle line data of the adjacent continuous sub-segments as the first target vehicle line data; If the number of data selection points contained in the continuous sub-segment adjacent to the continuous sub-segment before and / or after the continuous sub-segment is less than the learning result judgment threshold, the vehicle lane data stored in the segment in the map database is obtained as the first target vehicle lane data.

6. A lane positioning method according to claim 1, characterized in that: The method further comprises: Obtain the collection equipment information of each vehicle; The collection device information of each vehicle is associated and stored with the first target vehicle lane data of the vehicle obtained based on the second self-learning analysis result.

7. A lane positioning method, characterized in that: The method comprises: Collecting the collection device information, lane data and vehicle trajectory data of each vehicle within a preset range, wherein the lane data includes lane type and lane color; Performing cloud-based full-data self-learning analysis based on the acquired acquisition device information, vehicle line data, and vehicle trajectory data to obtain a third self-learning analysis result for each data selection point in the vehicle trajectory data; Connecting the third self-learning analysis results of the data selection points in series to obtain a fourth self-learning analysis result of a line segment of a corresponding length; Obtaining second target lane data based on the fourth self-learning analysis result; Obtaining second lane positioning data according to the second target vehicle lane data and current vehicle trajectory data; Get the lane data of the vehicle's current position; Comparing the vehicle lane data of the current position of the vehicle with the second target vehicle lane data to see whether they are consistent; If the two are consistent, the second lane positioning data is used as the output result; if the two are inconsistent, the lane data of the current position of the vehicle is corrected based on the fourth self-learning analysis result, and the second correction result is used as the output result; The cloud-based full data self-learning analysis is performed based on the acquired acquisition device information, vehicle line data, and vehicle trajectory data to obtain a third self-learning analysis result for each data selection point in the vehicle trajectory data, specifically: Count the second occurrence times of each vehicle lane data of each data selection point in the vehicle trajectory data; When it is determined that the second occurrence number of each of the vehicle lane data meets the third threshold, a second weight of each of the vehicle lane data at each data selection point is counted, where the second weight is a cumulative weight of each of the vehicle lane data; Determining whether the second weight satisfies a fourth threshold; If yes, the data selection point is learned successfully, and the lane data corresponding to the data selection point is used as the third self-learning analysis result; If not, the learning of the data selection point fails.

8. A lane positioning method according to claim 7, characterized in that: Before performing cloud-based full data self-learning analysis based on the acquired collection device information, vehicle lane data, and vehicle trajectory data, the method further includes: Determining whether the second occurrence number of each lane data satisfies a third threshold; If not, the lane data of each data selection point is saved as the second intermediate result.

9. A lane positioning method according to claim 7, characterized in that: The second weight of each lane data of each data selection point is specifically: Based on the vehicle line data of each data selection point in the vehicle trajectory data, the third occurrence number of each type of acquisition equipment is counted; Obtain the third weight corresponding to each type of acquisition equipment; Obtaining the weight of each type of acquisition equipment according to the product of the third number of outgoing lines of each type of acquisition equipment and the corresponding third weight; The sum of the weights of various types of acquisition equipment is calculated to obtain the second weight.

10. A lane positioning method according to claim 7, characterized in that: The third self-learning analysis result of the data selection point is connected in series to obtain a fourth self-learning analysis result of a line segment of corresponding length, specifically: The continuous data selection points having the same third self-learning analysis result are connected in series to obtain the learning results of each continuous sub-segment as the fourth self-learning analysis result.

11. A lane positioning method according to claim 10, characterized in that: The second target lane data is obtained based on the fourth self-learning analysis result, specifically: Determine the learning result judgment threshold according to the road type; When the number of data selection points contained in the continuous sub-segments is greater than or equal to the learning result judgment threshold; and If the learning result of the continuous sub-segment is successful, the vehicle lane data of the continuous sub-segment is used as the second target vehicle lane data; Or if the learning result of the continuous sub-segment is failure, the vehicle lane data stored in the segment in the map database is obtained as the second target vehicle lane data.

12. A lane positioning method according to claim 11, characterized in that: The obtaining of the second target lane data based on the fourth self-learning analysis result also includes: When the number of data selection points contained in the continuous sub-segments is less than the learning result judgment threshold; The number of data selection points contained in the continuous sub-segments adjacent to and before the continuous sub-segment is greater than or equal to the learning result judgment threshold, and the first half and the second half of the continuous sub-segment respectively use the vehicle line data of the adjacent continuous sub-segments as the second target vehicle line data; If the number of data selection points contained in the continuous sub-segment adjacent to the continuous sub-segment before and / or after the continuous sub-segment is less than the learning result judgment threshold, the vehicle lane data stored in the segment in the map database is obtained as the second target vehicle lane data.

13. A lane positioning method according to any one of claims 1 to 12, characterized in that: The lane data also includes the number of lanes, lane width, lane width, lane curvature and curvature change rate. Based on the number of lanes, lane width, lane width, lane curvature and curvature change rate of each data selection point obtained, a fifth self-learning analysis result is obtained by learning through a Kalman filter algorithm; The lane data also includes the positions of the stop line, zebra crossing, and road arrows. Based on the acquired longitude and latitude values ​​of the positions of the stop line, zebra crossing, and road arrows, a sixth self-learning analysis result is obtained by learning through an average value method; The lane data also includes road arrow directions for a straight line, left turn, U-turn, straight line plus U-turn, and straight line plus left turn. The number of each type of road arrow direction for each data selection point obtained is accumulated respectively, and the result with the largest number of times is counted as the seventh self-learning analysis result.

14. A lane positioning system, characterized in that: The system comprises: A first data collection module, used to collect lane data and vehicle trajectory data of each vehicle, wherein the lane data includes lane type and lane color; A first self-learning analysis module is used to perform vehicle terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data to obtain a first self-learning analysis result for each data selection point in the vehicle trajectory data; the vehicle terminal self-learning analysis based on the acquired vehicle lane data and vehicle trajectory data to obtain the first self-learning analysis result for each data selection point in the vehicle trajectory is specifically: counting the first occurrence number of each vehicle lane data at each data selection point in the vehicle trajectory data; when it is determined that the first occurrence number of each vehicle lane data meets the first threshold, counting the first weight of each vehicle lane data at each data selection point, the first weight being the occurrence probability of each vehicle lane data; and determining whether the first weight meets the second threshold; If yes, the data selection point is learned successfully, and the lane data corresponding to the data selection point is used as the first self-learning analysis result; if no, the data selection point is learned unsuccessfully; A second self-learning analysis module, used to connect the first self-learning analysis results of a preset number of data selection points in series to obtain a second self-learning analysis result of a line segment of a corresponding length; A first determination module, configured to obtain first target vehicle lane data based on the second self-learning analysis result; The first lane positioning module is used to obtain the first lane positioning data according to the first target vehicle lane data and the current vehicle trajectory data; obtain the vehicle lane data of the current position of the vehicle; compare the vehicle lane data of the current position of the vehicle with the first target vehicle lane data to see if they are consistent; if the two are consistent, use the first lane positioning data as the output result; if the two are inconsistent, correct the vehicle lane data of the current position of the vehicle based on the second self-learning analysis result, and use the first correction result as the output result.

15. A lane positioning system, characterized in that: The system comprises: A second data collection module is used to collect the collection device information, vehicle lane data and vehicle trajectory data of each vehicle within a preset range, wherein the vehicle lane data includes the vehicle lane type and vehicle lane color; A third self-learning analysis module is used to perform cloud-based full-data self-learning analysis based on the acquired acquisition device information, vehicle lane data and vehicle trajectory data to obtain a third self-learning analysis result for each data selection point in the vehicle trajectory data; the cloud-based full-data self-learning analysis based on the acquired acquisition device information, vehicle lane data and vehicle trajectory data to obtain the third self-learning analysis result for each data selection point in the vehicle trajectory data is specifically: counting the second occurrence number of each vehicle lane data of each data selection point in the vehicle trajectory data; when it is determined that the second occurrence number of each vehicle lane data meets the third threshold, counting the second weight of each vehicle lane data of each data selection point, the second weight being the cumulative weight of each vehicle lane data; determining whether the second weight meets the fourth threshold; if so, the data selection point is learned successfully, and the vehicle lane data corresponding to the data selection point is used as the third self-learning analysis result; if not, the data selection point fails to learn; A fourth self-learning analysis module, used for connecting the third self-learning analysis results of a preset number of data selection points in series to obtain a fourth self-learning analysis result of a line segment of a corresponding length; A second determination module, configured to obtain second target lane data based on the fourth self-learning analysis result; The second lane positioning module is used to obtain the second lane positioning data according to the second target vehicle lane data and the current vehicle trajectory data; obtain the vehicle lane data of the current position of the vehicle; compare the vehicle lane data of the current position of the vehicle with the second target vehicle lane data to see if they are consistent; if the two are consistent, use the second lane positioning data as the output result; if the two are inconsistent, correct the vehicle lane data of the current position of the vehicle based on the fourth self-learning analysis result, and use the second correction result as the output result.

16. An electronic device, characterized in that: The electronic device comprises: at least one processor and at least one memory; the memory is used to store one or more program instructions; The processor is used to run one or more program instructions to execute a lane positioning method as described in any one of claims 1-13.

Images