Crowdsourcing map special lane updating method, device and equipment and storage medium

Abstract

The invention relates to the technical field of high-precision maps, and discloses a crowdsourcing map special lane updating method, device and equipment and a storage medium, and the method comprises the steps: obtaining a first road picture shot in a vehicle driving process and the track information of a vehicle; and according to the track information, matching the first road picture with a lane center line in crowdsourcing map data to obtain a second road picture bound with the lane center line. And identifying the special lane features in the second road picture, and generating a change point object containing special lane identification information. And according to the change point object, carrying out differential updating on the special lane data in the crowdsourcing map. According to the invention, poor time efficiency and identification errors caused by depending on a professional acquisition vehicle and a manual operation mode are avoided, low-efficiency operation of repeated updating and coverage of the whole special lane data is avoided, and high-efficiency and high-precision updating and maintenance of the special lane data can be realized.

Description

Technical Field

[0001] The present invention relates to the field of high-precision map technology, and in particular to a method, device, equipment and storage medium for updating dedicated lanes on a crowdsourcing map. Background Art

[0002] Self-driving cars need a high-precision, high-dimensional, and high-timeliness electronic map to achieve accurate positioning, navigation, and control functions. Crowdsourcing maps can help self-driving cars make reasonable driving decisions by collecting road information and making advance predictions about the vehicle's surrounding environment. Among them, dedicated lane information is an important part of crowdsourcing maps, which has a direct impact on the safety and efficiency of self-driving cars. Dedicated lanes refer to lanes that only allow specific types of vehicles to travel within a specific time and range, such as bus lanes, motor vehicle lanes, and non-motor vehicle lanes.

[0003] At present, the production and update of crowdsourced map dedicated lane data mainly rely on professional collection vehicles and manual operations. This method has the defects of low efficiency and low accuracy when updating dedicated lane data, and it is difficult to meet the requirements of autonomous driving for high precision and high timeliness of crowdsourced map dedicated lane data. Summary of the invention

[0004] In order to overcome the defects of low efficiency and low accuracy in updating crowdsourced map dedicated lane data in the prior art, the present invention proposes the following technical solutions:

[0005] In a first aspect, the present invention proposes a crowdsourcing map dedicated lane update method, comprising:

[0006] A first road image taken during the driving of the vehicle and the trajectory information of the vehicle are obtained.

[0007] According to the trajectory information, the first road image and the lane centerline in the crowdsourced map data are matched to obtain a second road image bound to the lane centerline.

[0008] The dedicated lane features in the second road image are identified to generate a change point object including dedicated lane identification information.

[0009] According to the change point object, the dedicated lane data in the crowdsourced map is differentially updated.

[0010] In a second aspect, the present invention further proposes a crowdsourcing map dedicated lane update method and device, comprising:

[0011] The acquisition module is used to acquire the first road picture taken during the vehicle's driving and the vehicle's trajectory information.

[0012] A matching module is used to match the first road image with the lane center line in the crowdsourced map data according to the trajectory information to obtain a second road image bound to the lane center line.

[0013] The recognition module is used to recognize the special lane features in the second road image and generate a change point object containing the special lane identification information.

[0014] An updating module is used to perform differential update on the dedicated lane data in the crowdsourced map according to the change point object.

[0015] In a third aspect, the present invention further proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the operations performed by the crowdsourcing map dedicated lane update method as described in the first aspect are implemented.

[0016] In a fourth aspect, the present invention further proposes a computer-readable storage medium having a program stored thereon, wherein the program is executed by a processor to perform the operations performed by the crowdsourcing map dedicated lane updating method described in the first aspect.

[0017] The beneficial effects of the present invention include at least:

[0018] (1) The present invention utilizes a first road image taken during vehicle travel and the vehicle's trajectory information, and matches the first road image with the lane centerline in crowdsourced map data, thereby avoiding the timeliness and recognition errors caused by relying on professional data collection vehicles and manual operations, and can obtain the latest status of the dedicated lane in real time, thereby improving the efficiency and accuracy of updating the dedicated lane data.

[0019] (2) The present invention identifies the dedicated lane features in the second road image, generates a change point object containing the dedicated lane identification information, and differentially updates the dedicated lane data in the crowdsourced map, which can effectively extract the key points that have changed in the dedicated lane data, avoid the inefficient operation of repeatedly updating and covering the entire dedicated lane data, and achieve high-efficiency and high-precision updating and maintenance of the dedicated lane data. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flowchart of a crowdsourcing map dedicated lane updating method provided in an embodiment of the present application.

[0021] Figure 2 This is a schematic diagram of the principle of predicting poor-quality images from trajectory object features in an embodiment of the present application.

[0022] Figure 3This is a schematic diagram of a process for performing quality scoring on a first road image in an embodiment of the present application.

[0023] Figure 4 This is a schematic diagram of the extended update of dedicated lane data in an embodiment of the present application.

[0024] Figure 5 This is an architectural diagram of the crowdsourcing map dedicated lane updating device provided in an embodiment of the present application.

[0025] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred technical solutions. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred technical solutions are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0027] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0028] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0029] Specifically, Figure 1 A flowchart of a crowdsourcing map dedicated lane updating method provided in an embodiment of the present application.

[0030] like Figure 1 As shown, the crowdsourcing map dedicated lane updating method includes the following steps:

[0031] S10: Acquire a first road image taken during the driving of the vehicle and the trajectory information of the vehicle.

[0032] In the specific implementation process, a vehicle equipped with a forward-looking acquisition device is used to collect the first road image and trajectory information on the road while driving, and the data is uploaded to the cloud for subsequent processing.

[0033] Optionally, in one embodiment of the present application, after acquiring the first road image, the images containing shadows, bad pixels, tunnel scenes, and upper and lower layer scenes, as well as images in which the track has a horizontal or vertical distance offset of more than 5 meters in the first road image are filtered.

[0034] Optionally, in an embodiment of the present application, after acquiring the first road picture, performing a quality score on the first road picture, the specific steps include:

[0035] S101: performing semantic segmentation on the first road image to obtain an occlusion area size and a road surface information ratio of the first road image.

[0036] S102: According to the size of the blocked area and the proportion of the road surface information, the first road image is scored using a preset rule to obtain a first scoring result.

[0037] S103: Using preset rules to train and learn features of several types of objects in the image classifier to obtain a trained image classifier.

[0038] S104: Using the trained image classifier to classify and score the first road image to obtain a second scoring result.

[0039] S105: Perform a weighted summation on the first scoring result and the second scoring result to obtain a quality score of the first road picture.

[0040] It is understandable that since the crowd-source acquisition equipment (crowd-source box) is not a professional video recorder, it is greatly affected by the external environment. In addition, user behavior is uncontrollable, the equipment is not placed in a fixed position, and driving is not standardized, resulting in less available information in the generated images and greater difficulty in recognition. Therefore, poor quality images are predicted based on trajectory features, and images of unreliable quality are cleared or evaluated before subsequent processing.

[0041] Optionally, in one embodiment of the present application, meteorological parameters and lighting parameters of the vehicle driving process are acquired in real time, and the solar altitude angle is calculated, and the meteorological parameters, lighting parameters and solar altitude angle are used as object features of the classifier for training and learning.

[0042] It is understandable that due to the different degrees of obstruction and influence of rain and snow on ground information and collection equipment, the intensity of sunlight can easily cause the equipment to produce overexposed or dark photos. The solar altitude angle needs to be calculated through trajectory features (coordinate information / longitude and latitude information). The calculation formula of the solar altitude angle is as follows:

[0043] sinh=sinasinb+cosacosbcost

[0044] Among them, a represents the solar declination, b represents the geographical latitude of the observation site (both the solar declination and the geographical latitude are positive with the north latitude and negative with the south latitude), and t represents the local time.

[0045] Meteorological conditions and lighting conditions are mainly obtained through real-time recorded weather information. As one of the training object features of the classifier, the corresponding information is stored in the trajectory point file bound to the image for subsequent reading.

[0046] In the specific implementation process, Figure 2 and Figure 3 As shown, for clear quality problems, such as occlusion and tilt caused by improper placement of equipment, semantic segmentation is introduced to quantify and score the pictures. In this embodiment, due to the tilt and occlusion caused by improper placement, the score can be based on the size of the occlusion block in the image and the proportion of road surface information in the tilted image. If the occlusion area is greater than 50% or the proportion of road surface information in the tilted image is less than 50%, the first score result is 0; if the occlusion area is less than 50% or the proportion of road surface information in the tilted image is greater than 50%, the first score result value is increased according to the proportion interval (i.e., the occlusion area is 40%-50% or the road surface information accounts for 50%-60%, the first score result value is 10 points, and so on), and the maximum score is no more than 50 points.

[0047] Table 1 Characteristics of 22 types of objects

[0048] Vehicle gland Qualified pictures Upward Snow on the ground facing inwards Block from above Too dark Wipers Rotate left and right Invert Right side occlusion Backlight Off-road Black and gray tilt Vague Downward Hood rain Reflective tunnel Abnormal picture

[0049] For the blur and backlight caused by the shaking of equipment due to road bumps, a fine image evaluation criterion is established through the 22-category feature image classifier of the computer vision image classification model, as shown in Table 1. The 22-category feature image classifier fully considers deformation, lighting / illumination conditions, background clutter, viewpoint changes, and scale changes, trains and learns 22 object features, and divides different images into different types. Except for qualified images with a score of 50, the scores of other types are all 0, and the second scoring results of each image are obtained.

[0050] Finally, a weighted sum is performed on the first scoring result and the second scoring result to obtain a quality score of the first road image.

[0051] S20: Matching the first road image with a lane centerline in the crowdsourced map data according to the trajectory information to obtain a second road image bound to the lane centerline.

[0052] In the specific implementation process, the lane centerline closest to the first road image binding trajectory in the crowdsourced map data is bound according to the nearest neighbor matching distance through the trajectory information (spatial position and geometric topology information) of the first road image binding trajectory, and recorded in the associated lane centerline field in the trajectory point.

[0053] In addition, for the middle section of a regular road (no upper and lower layers, non-intersections), the trajectory points are bound by searching for the nearest lane centerline through the left and right 1.5m spatial buffer. Since the current high-precision map industry standard does not make special lanes within the intersection range, the upper and lower layer scenes do not have elevation information on the trajectory points and pictures, and it is impossible to determine the centerline to be bound only by binding the road information. Subsequent updates to special lanes are mainly made by extending the front and rear lanes. Therefore, the pictures within the intersection range and the upper and lower layers cannot be bound to the centerline pictures and trajectories at this stage. Among them, the principle of distinguishing intersections from upper and lower layer scenes is mainly that the image binding trajectory is consistent with the distance of 2 or more center lines, and the nearest and only center line cannot be selected. The regular scene is only closest to 1 center line.

[0054] It is understandable that by binding the lane centerline closest to the bound trajectory of the first road image in the crowdsourced map data and recording it in the associated lane centerline field in the trajectory point, the lane position of the first road image can be determined for subsequent image recognition and differential update. The trajectory point corresponding to the first road image refers to the trajectory information recorded by the device at the same time when the image is collected, including coordinates, time, speed, etc. The trajectory point is the spatial location information of the image, which can be used to calculate geographic parameters and bind lane centerlines.

[0055] S30: Identify the features of the dedicated lane in the second road image, and generate a change point object containing dedicated lane identification information.

[0056] Optionally, in one embodiment of the present application, the specific step of identifying the dedicated lane feature in the second road image includes:

[0057] S301: Acquire at least one dedicated lane feature including ground text, aerial signboard or special line type from the second road image as an identification object.

[0058] S302: Calculate the coordinates of the center point of the identified object.

[0059] S303: Mapping the center point coordinates to a trajectory coordinate system bound to the second road image, where the trajectory coordinate system is a coordinate system corresponding to trajectory information recorded during the vehicle's driving process.

[0060] S304: Calculate the trajectory point object closest to the center point coordinates in the trajectory coordinate system, and generate a change point object at the position of the trajectory point object.

[0061] In this embodiment, the recognition objects include ground text (bus-only, bus lane, BRT), aerial signs (bus icon + bus-only, bus line dedicated lane), special line types (yellow dashed line), and a second road image only generates one change point. The coordinates of the change point are taken from the coordinates of the trajectory point bound to the image, that is, a point object is generated at the corresponding trajectory point position.

[0062] Optionally, in one embodiment of the present application, generating a change point object containing dedicated lane identification information specifically includes: storing the corresponding recognition object, the quality score of the first road image, and the lane center line bound to the second road image in the trajectory point object to obtain the change point object.

[0063] In this embodiment, the change point object only represents the information that there is a dedicated lane on the ground. The change point object directly inherits the image binding trajectory information, and records the associated center line, dedicated lane type, recognition object classification (ground text / aerial sign / special marking) and quality assessment information; if there are multiple recognition objects in the image, the object type information is stored in the recognition object classification, that is, if there are three types of objects, namely ground text, aerial sign and special marking in the image at the same time, the recognition object classification field of the change point needs to record "ground text / aerial sign / special marking".

[0064] Optionally, in one embodiment of the present application, when generating a change point object containing dedicated lane identification information, the change point object needs to be deduplicated. A plurality of spatial buffers are generated along the path direction of the lane centerline at a preset distance; for each of the spatial buffers, if there are two or more change point objects, the change point object with the highest quality score of the associated first road image is retained; if the quality scores of the associated first road images are the same, the change point object with a larger number of identification object information field storages is retained; if there are identification object information fields with the same number of storages, a change point object is randomly selected for retention.

[0065] In the specific implementation process, since the dedicated lane type data is updated at the road section level (intersection-intersection middle range), in order to avoid information interference and reduce the amount of data, only one change point within a certain range of the same lane is retained as traction positioning, and the change points need to be deduplicated, mainly through spatial position (coordinate point position) and the center line of the associated lane. The change points associated with the center line of the same lane are deduplicated through a 10m spatial buffer. If there are other change points within a 10m range, the change points with low quality assessment scores are excluded, that is, there are 2 change points within a 10m range, one change point image score is 40, and the other change point image score is 20. If the score is 60, the change point with a score of 40 is removed; if the scores are the same, the change point with richer lane type information is retained, that is, the point with a larger number of identification type classification fields stored in the change point; if there are two change points, one of which has only "air sign" in the identification type classification, and the second change point has "ground text / space sign" in the identification type classification, then the second change point is retained; if the identification type information is consistent, one is randomly retained; the spatial buffer zone does not cross the intersection for judgment, that is, if the spatial buffer zone of the change point overlaps with the stop line / crosswalk, only the change point between the change point and the stop line / crosswalk is deleted, and it is not extended forward.

[0066] S40: performing differential update on the dedicated lane data in the crowdsourced map according to the change point object.

[0067] In this embodiment, the actual dedicated lane and its restriction information contained in the change point object are parsed from the change point object; if the lane centerline associated with the change point object lacks the actual dedicated lane, the actual dedicated lane and its restriction information are added to the corresponding position of the crowdsourcing map; if other dedicated lanes that are inconsistent with the actual dedicated lane exist in the lane centerline associated with the change point object, the other dedicated lanes are changed to ordinary lanes at the corresponding position of the crowdsourcing map.

[0068] In the specific implementation process, the change points after deduplication are bound to the lane centerline ID and the lane center in the high-precision map data to perform field information difference. The difference content is whether the lane type of the centerline associated with the change point includes the special lane type and whether the restriction information is correct. That is, the updated content is mainly the omission of special lanes that do not exist in the actual data and the redundant production of special lanes that do exist in the actual data. The final change point output after the difference represents the real change point where the data is inconsistent with the actual data, that is, the point information inconsistent with the high-precision map data. Omission production requires the addition of special lanes in the centerline lane type. When the special lane type conflicts with other lane types, they cannot coexist. Therefore, only the special lane is retained (that is, if it is determined that the special lane type is omitted, the existing lane type on the high-precision map data line needs to be deleted), and the restriction information on the lane centerline is added synchronously (the restriction information is a field of the lane centerline object, and the addition is mainly done by filling the text information "car" into the field). The restriction object defaults to a car, such as Figure 4 As shown, the newly added special lane type needs to be extended and judged. If the center line of the lane associated with the final change point is in the road section, the front and rear center lines of the same road section range need to maintain the center line type as a special lane (within the intersection-intersection range) and end at the intersection. If there are change points after the intersection, continue to extend. If not, end before the intersection. Redundant production requires deleting the special lane type in the center line lane type, changing it to a normal lane, and deleting the restriction information. When all change points are traversed, the update of the special lane data in the crowdsourcing map is completed.

[0069] It can be understood that the crowdsourced map dedicated lane update method proposed in this embodiment uses the first road picture taken during the vehicle's driving and the vehicle's trajectory information, and matches the first road picture with the lane centerline in the crowdsourced map data, avoiding the timeliness and recognition errors caused by relying on professional collection vehicles and manual operations, and can obtain the latest status of the dedicated lane in real time, improving the efficiency and accuracy of the dedicated lane data update. The dedicated lane features in the second road picture are identified, and a change point object containing the dedicated lane identification information is generated and the dedicated lane data in the crowdsourced map is differentially updated, which can effectively extract the key points that have changed in the dedicated lane data, avoid the inefficient operation of repeatedly updating and covering the entire dedicated lane data, and can achieve efficient and high-precision update and maintenance of the dedicated lane data.

[0070] Secondly, the crowdsourcing map dedicated lane updating device proposed according to the embodiment of the present application is described with reference to the accompanying drawings.

[0071] Figure 5 It is an architectural diagram of a crowdsourcing map dedicated lane updating device according to an embodiment of the present application.

[0072] like Figure 5As shown, the updating device includes: an acquisition module 100 , a matching module 200 , an identification module 300 and an updating module 400 .

[0073] Among them, the acquisition module 100 is used to obtain a first road picture taken during the driving process of the vehicle and the trajectory information of the vehicle; the matching module 200 is used to match the first road picture and the lane center line in the crowdsourced map data according to the trajectory information, and obtain a second road picture bound to the lane center line; the identification module 300 is used to identify the dedicated lane features in the second road picture and generate a change point object containing the dedicated lane identification information; the update module 400 is used to differentially update the dedicated lane data in the crowdsourced map according to the change point object.

[0074] It should be noted that the aforementioned explanation of the crowd-sourced map dedicated lane updating method embodiment is also applicable to the crowd-sourced map dedicated lane updating device of this embodiment, and will not be repeated here.

[0075] Figure 6 The computer device 500 provided in the embodiment of the present application is a schematic diagram of the structure of the computer device 500. The computer device 500 includes: a memory 501, a processor 502, and a computer program stored in the memory 501 and executable on the processor 502.

[0076] When the processor 502 executes the program, the crowdsourcing map dedicated lane updating method provided in the above embodiment is implemented.

[0077] Furthermore, the computer device 500 also includes: a communication interface 503 for communication between the memory 501 and the processor 502 .

[0078] The memory 501 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0079] If the memory 501, the processor 502 and the communication interface 503 are implemented independently, the communication interface 503, the memory 501 and the processor 502 can be connected to each other through a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0080] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 503 are integrated on a chip, the memory 501, the processor 502 and the communication interface 503 can communicate with each other through an internal interface.

[0081] The processor 502 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0082] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above crowdsourcing map dedicated lane update method.

[0083] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0084] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0085] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0086] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array, a field programmable gate array, etc.

[0087] A person skilled in the art may understand that all or part of the steps in the above-mentioned embodiment method may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0088] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A crowdsourcing map dedicated lane update method, characterized in that: include: Acquire a first road image taken during the driving of the vehicle and the trajectory information of the vehicle; According to the trajectory information, matching the first road image with a lane centerline in the crowdsourced map data to obtain a second road image bound to the lane centerline; Identify the dedicated lane features in the second road image and generate a change point object containing dedicated lane identification information; According to the change point object, the dedicated lane data in the crowdsourced map is differentially updated.

2. The crowdsourcing map dedicated lane updating method according to claim 1, characterized in that: After acquiring the first road picture, the method further includes: Performing semantic segmentation on the first road image to obtain an occlusion area size and a road surface information ratio of the first road image; Scoring the first road image according to the size of the blocked area and the proportion of the road surface information using a preset rule to obtain a first scoring result; Using preset rules to train and learn features of several types of objects in the image classifier, a trained image classifier is obtained; Using the trained image classifier to classify and score the first road image to obtain a second scoring result; A weighted sum is performed on the first scoring result and the second scoring result to obtain a quality score of the first road image.

3. The crowdsourcing map dedicated lane updating method according to claim 2, characterized in that: The method also includes acquiring meteorological parameters and lighting parameters during the vehicle driving process in real time, calculating the solar altitude angle, and using the meteorological parameters, lighting parameters and solar altitude angle as object features of the classifier for training and learning.

4. The crowdsourcing map dedicated lane updating method according to claim 1, characterized in that: Performing image recognition on the dedicated lane features in the second road image specifically includes: Acquire at least one dedicated lane feature including ground text, aerial sign or special line type as a recognition object from the second road image; Calculating the coordinates of the center point of the identified object; Mapping the center point coordinates to a trajectory coordinate system bound to the second road image, where the trajectory coordinate system is a coordinate system corresponding to the trajectory information recorded during the vehicle's travel; A trajectory point object having the shortest coordinate distance to the center point is calculated in the trajectory coordinate system, and a change point object is generated at the position of the trajectory point object.

5. The crowdsourcing map dedicated lane updating method according to claim 1, 2 or 4, characterized in that: Generate a change point object containing the dedicated lane identification information, including: The corresponding recognition object, the quality score of the first road image, and the lane center line bound to the second road image are stored in a track point object to obtain a change point object.

6. The crowd-sourced map dedicated lane updating method according to claim 5, characterized in that: After generating the change point object including the dedicated lane identification information, the method further includes: Generate several spatial buffers along the path direction of the lane centerline according to the preset distance; For each of the spatial buffers, if there are two or more change point objects, the change point object with the highest quality score of the associated first road image is retained; if the quality scores of the associated first road images are the same, the change point objects with a larger number of identification object information fields are retained; if there are identification object information fields with the same number of storages, a change point object is randomly selected for retention.

7. The crowd-sourced map dedicated lane updating method according to claim 1, characterized in that: According to the change point object, the dedicated lane data in the crowdsourced map is differentially updated, specifically including: Parsing the change point object to obtain the on-site dedicated lane and its restriction information contained therein; If the lane centerline associated with the change point object lacks the actual dedicated lane, the actual dedicated lane and its restriction information are added at the corresponding position on the crowdsourced map; if other dedicated lanes that are inconsistent with the actual dedicated lane exist in the lane centerline associated with the change point object, the other dedicated lanes are changed to ordinary lanes at the corresponding position on the crowdsourced map.

8. The crowdsourcing map dedicated lane updating method according to claim 1, characterized in that: After acquiring the first road image, the method further includes filtering images containing shadows, bad pixels, tunnel scenes, and upper and lower layer scenes in the first road image, as well as images in which the track has a horizontal or vertical distance offset of more than 5 meters.

9. A crowd-sourced map dedicated lane updating device, characterized in that: include: An acquisition module, used to acquire a first road image taken during the driving of the vehicle and the trajectory information of the vehicle; A matching module, configured to match the first road image with a lane centerline in the crowdsourced map data according to the trajectory information to obtain a second road image bound to the lane centerline; A recognition module, used to recognize the features of the dedicated lane in the second road image and generate a change point object containing the identification information of the dedicated lane; An updating module is used to perform differential update on the dedicated lane data in the crowdsourced map according to the change point object.

10. An electronic device, characterized in that: The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the operations performed by the crowdsourcing map dedicated lane updating method as described in any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, and the program is used by a processor to execute the operations performed by the crowdsourcing map dedicated lane updating method as described in any one of claims 1 to 8.

Images