Intelligent automobile map data auditing method and device, equipment and medium

Abstract

The invention relates to an intelligent automobile map data auditing method. The method comprises the following steps: acquiring geographic space data acquired by each vehicle-mounted sensor of an intelligent automobile; performing point cloud target detection on the laser point cloud data based on a preset point cloud target detection model, and determining a position corresponding to a first map data sensitive area in the laser point cloud data; performing target detection on the image data based on a preset image target detection model, and determining a position corresponding to a second map data sensitive area in the image data; based on the position corresponding to the first map data sensitive area and the position corresponding to the second map data sensitive area, a to-be-audited sensitive information set of the geographic space data is determined, and sensitive information in the geographic space data is determined according to the to-be-audited sensitive information set; and filtering sensitive information in the geographic space data according to a preset sensitive information filtering algorithm. According to the invention, convenient and efficient map review work management can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a smart car map data auditing method and device, electronic equipment and computer readable storage medium. BACKGROUND

[0003] With the rapid development of the automatic driving industry, networked intelligent vehicle sensors are everywhere. The behaviors of geographic spatio-temporal data collection, processing, mapping, transmission and storage bring great challenges to the automatic driving map review due to their scene characteristics such as hardware platform, intelligent computing process, communication environment, storage and regulatory review process. Due to the lack of intelligent map review technology and software and platform support, map review work relies heavily on manual interpretation and visual interpretation by experienced map review experts to find and evaluate the compliance of the map, thus becoming a highly labor-intensive work. It has the outstanding problems of high professional requirement, strong subjectivity, low efficiency and high cost, and it is difficult to meet the all-weather, large-scale and large-batch map review demand.

[0004] At present, the map review scheme for a single / concrete map data model cannot be widely applied and adapted, and can only review part of the calculation links of map collection and processing, including spatial position encryption, road element compliance, sensitive place index, sensitive information retrieval of result image, and map review for automatic driving map product data, but often focuses on the post-processing results of smart car map data, i.e. map data products. The data collection, transmission and preprocessing process of the smart car map data sensor lacks supervision and cannot realize the whole process of smart car geographic spatial data review and monitoring. SUMMARY

[0005] The present application aims to solve the above problems and provide a smart car map data auditing method, corresponding device, electronic equipment and computer readable storage medium.

[0006] To meet the various purposes of the present application, the present application adopts the following technical solutions:

[0007] A smart car map data auditing method is proposed to adapt to one of the purposes of the present application, comprising:

[0008] In response to a smart car map data auditing instruction, geographic spatial data collected by each vehicle-mounted sensor of a smart car is obtained, wherein the geographic spatial data includes image data and laser point cloud data;

[0009] Point cloud target detection is performed on the laser point cloud data to determine the position corresponding to the first map data sensitive area in the laser point cloud data;

[0010] perform target detection on the image data to determine a position corresponding to the second map data sensitive region in the image data;

[0011] determine a set of sensitive information to be audited in the geospatial data based on the position corresponding to the first map data sensitive region in the laser point cloud data and the position corresponding to the second map data sensitive region in the image data, and determine the sensitive information in the geospatial data according to the set of sensitive information to be audited;

[0012] filter the sensitive information in the geospatial data to complete the auditing of the intelligent vehicle map data.

[0013] Optionally, before the step of obtaining the geospatial data collected by each vehicle-mounted sensor of the intelligent vehicle, the method comprises:

[0014] detecting whether the geospatial data collected by each vehicle-mounted sensor meets a preset serial communication type and a sensor data standard, and if so, inputting the geospatial data collected by each vehicle-mounted sensor into the intelligent vehicle geospatial data auditing system for processing.

[0015] Optionally, the step of performing point cloud target detection on the laser point cloud data to determine the position corresponding to the first map data sensitive region in the laser point cloud data comprises:

[0016] when detecting that the distance between the current position of the intelligent vehicle and the position of the sensitive information is less than a first preset distance, calling a preset point cloud target detection model to perform point cloud target detection on the laser point cloud data to determine the position corresponding to the first map data sensitive region in the laser point cloud data.

[0017] Optionally, the step of performing target detection on the image data to determine the position corresponding to the second map data sensitive region in the image data comprises:

[0018] when detecting that the distance between the current position of the intelligent vehicle and the position of the sensitive information is less than a second preset distance, or the set of sensitive picture information is not empty, or the set of sensitive character strings is not empty, performing target detection on the image data based on a preset image target detection model to determine the position corresponding to the second map data sensitive region in the image data.

[0019] Optionally, the step of determining a set of sensitive information to be audited in the geospatial data based on the position corresponding to the first map data sensitive region in the laser point cloud data and the position corresponding to the second map data sensitive region in the image data, and determining the sensitive information in the geospatial data according to the set of sensitive information to be audited comprises:

[0020] According to the position corresponding to the first map data sensitive region in the laser point cloud data and the position corresponding to the second map data sensitive region in the image data, determine the sensitive information that needs to be focused on in the current stage in the sensitive geographic information database, to determine a set of sensitive information to be audited M;

[0021] Read a new information record from the set of sensitive information to be audited, denoted as sensitive information to be audited record a;

[0022] Perform image retrieval analysis on the geographic space data of the current frame, and calculate the matching with the sensitive information to be audited record a, to determine whether the geographic space data of the current frame contains sensitive information associated with the sensitive information to be audited record a;

[0023] If the geographic space data of the current frame contains sensitive information associated with the sensitive information to be audited record a, determine the type of the sensitive information to be audited record a and its coordinate position in the geographic space data of the current frame, and add it to the set of sensitive information detection results N.

[0024] Optionally, the step of filtering the sensitive information in the geographic space data to complete the auditing of the intelligent vehicle map data comprises:

[0025] Obtain a sensitive information detection result record in the set of sensitive information detection results N, wherein the sensitive information detection result record comprises a sensitive information type and a sensitive information region;

[0026] Determine a preset expansion radius, and expand a circle with any point O(i, j) in the sensitive information region as the center to the four directions according to the preset expansion radius;

[0027] Remove the intersection area of the sensitive information region and the circle to filter the sensitive information in the geographic space data.

[0028] Optionally, the sensitive information comprises one or any combination of a sensitive geographic landmark picture, a sensitive administrative planning picture, a sensitive geographic landmark identifier, and a sensitive administrative planning identifier.

[0029] Another object of the present application is to provide an intelligent vehicle map data auditing device, comprising:

[0030] A sensor data acquisition module is configured to acquire geographic space data collected by each vehicle-mounted sensor of an intelligent vehicle in response to an intelligent vehicle map data auditing instruction, wherein the geographic space data comprises image data and laser point cloud data;

[0031] The first sensitive area determination module is configured to perform point cloud target detection on the laser point cloud data based on a preset point cloud target detection model, and determine a position corresponding to a first map data sensitive area in the laser point cloud data;

[0032] The second sensitive area determination module is configured to perform target detection on the image data based on a preset image target detection model, and determine a position corresponding to a second map data sensitive area in the image data;

[0033] The sensitive information determination module is configured to determine a set of sensitive information to be audited of the geographic space data based on the position corresponding to the first map data sensitive area in the laser point cloud data and the position corresponding to the second map data sensitive area in the image data, and determine sensitive information in the geographic space data according to the set of sensitive information to be audited.

[0034] The map data auditing module is configured to filter the sensitive information in the geographic space data according to a preset sensitive information filtering algorithm, so as to complete the auditing of the intelligent vehicle map data.

[0035] Another object of the present application is to provide an electronic device comprising a central processing unit and a memory, wherein the central processing unit is configured to invoke and run a computer program stored in the memory to perform the steps of the intelligent vehicle map data auditing method.

[0036] Another object of the present application is to provide a computer readable storage medium storing a computer program implemented according to the intelligent vehicle map data auditing method in the form of computer readable instructions, wherein the computer program is invoked and run by a computer to perform the steps included in the corresponding method.

[0037] Compared with the prior art, the present application is directed to the map review of the finished product data of the automatic driving map in the prior art, but often focuses on the post-processing results of the intelligent vehicle map data. The data collection, transmission and preprocessing process of the intelligent vehicle map data sensor lack supervision, and the full-process intelligent vehicle geographic space data review and monitoring cannot be realized. The present application includes but is not limited to the following beneficial effects:

[0038] Firstly, the intelligent vehicle map data auditing method of the present application proposes a real-time map data intelligent auditing method based on edge intelligent computing, based on the feature analysis of intelligent vehicle geographic spatial data sensors and data processing, and the map auditing work target. Through the embedded computing mode, the map data sensors of the intelligent vehicle are adaptively adapted, so as to monitor the intelligent vehicle geographic spatial data acquisition process, and through the image intelligent algorithm such as image target detection and tracking, data mining, the sensitive targets of the image are retrieved and the image is repaired. Through data classification and feature extraction of the sensitive target information library, high-accuracy and low-time-consumption sensitive information extraction and omission and inconsistency in the map data are realized;

[0039] Secondly, the intelligent vehicle map data auditing method of the present application, based on the real-time auditing-oriented automatic driving map data auditing system architecture, is loosely coupled with the intelligent vehicle geographic spatial data acquisition, sensitive map element identification and data repair, so as to realize convenient and efficient map auditing work management and improve the executability and interpretability of the map auditing process.

[0040] Thirdly, the intelligent vehicle map data auditing method of the present application, based on the data classification and feature extraction of the sensitive target information library, reduces the requirements for the hardware platform storage and computing performance of the map auditing intelligent computing, and at the same time, the requirements of the map auditing policy and regulatory agencies are more convenient to respond. BRIEF DESCRIPTION OF DRAWINGS

[0041] The above and / or additional aspects and advantages of the present application will become apparent and more readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:

[0042] Figure 1 An exemplary network architecture adopted by the intelligent vehicle geographic spatial data auditing system of the present application;

[0043] Figure 2 A flowchart of the intelligent vehicle map data auditing method in the embodiment of the present application;

[0044] Figure 3 A schematic diagram of the registration and registration management of geographic spatial data according to serial communication types and data standard protocols in the embodiment of the present application;

[0045] Figure 4 A flowchart of detecting sensitive geographic information in the embodiment of the present application;

[0046] Figure 5 A flowchart of removing sensitive information in the embodiment of the present application;

[0047] Figure 6 A schematic diagram of calculating the RGB value of the sensitive information area in the embodiment of the present application;

[0048] Figure 7 This is a schematic block diagram of the intelligent vehicle map data verification device in the embodiments of this application;

[0049] Figure 8 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0050] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0051] Those skilled in the art will understand that although the various methods in this application are described based on the same concept and thus present commonality among them, they can be performed independently unless otherwise specified. Similarly, the various embodiments disclosed in this application are all based on the same inventive concept; therefore, concepts expressed in the same way, as well as concepts that are appropriately changed for convenience but are expressed differently, should be understood equivalently.

[0052] Unless otherwise expressly stated, the various embodiments disclosed in this application can be combined in a cross-cutting manner to flexibly construct new embodiments, as long as such combination does not depart from the inventive spirit of this application and can meet the needs of the prior art or solve a certain deficiency in the prior art. Those skilled in the art should be aware of such modifications.

[0053] Please see Figure 1 An exemplary network architecture for an intelligent vehicle geospatial data review system is as follows: Figure 1 As shown, the intelligent vehicle map data review method of this application can be implemented based on the intelligent vehicle geospatial data review system. The overall framework of the intelligent vehicle geospatial data review system mainly includes two parts: the vehicle sensor data acquisition subsystem and the map data encryption subsystem.

[0054] In some embodiments, compared to traditional sensor data acquisition, the additional work involved in vehicle-mounted sensor data acquisition primarily concerns access management for vehicle-mounted sensors. To conduct comprehensive review of intelligent vehicle map data, necessary registration and filing are required for vehicle-mounted map data sensors (such as cameras, LiDAR, millimeter-wave radar, GNSS positioning, etc.). Before data acquisition, serial communication type detection and adaptation, sensor data standard and protocol adaptation, and other matters must be performed.

[0055] In some embodiments, the map data encryption subsystem is a main part of the automobile geospatial data review, mainly including: a map data sensitive area judgment module, a sensitive geographic information detection module, a sensitive information area filtering and completion module, and a sensitive information filtering module, etc.

[0056] The map review authority can provide necessary information infrastructure for commercial automobile map data collection and application by designing, establishing and publishing a sensitive geographic information database.

[0057] Based on the above exemplary scenarios, please refer to Figure 2 In one embodiment of the intelligent automobile map data review method of the present application, the method comprises:

[0058] Step S10, in response to the intelligent automobile map data review instruction, acquiring the geospatial data collected by each vehicle-mounted sensor of the intelligent automobile, wherein the geospatial data at least includes image data and laser point cloud data;

[0059] The intelligent automobile geospatial data review system can respond to the intelligent automobile map data review instruction to acquire the geospatial data collected by each vehicle-mounted sensor of the intelligent automobile, wherein the geospatial data includes GNSS data, image data and laser point cloud data, etc.

[0060] In some embodiments, before the step of acquiring the geospatial data collected by each vehicle-mounted sensor of the intelligent automobile, the method comprises:

[0061] Detecting whether the geospatial data collected by each vehicle-mounted sensor meets the pre-set serial communication type and sensor data standard, and if so, inputting the geospatial data collected by each vehicle-mounted sensor into the intelligent automobile geospatial data review system for processing.

[0062] Specifically, please refer to Figure 3 Compared with the traditional sensor data collection, the additional work of vehicle-mounted sensor data collection mainly involves the access management of vehicle-mounted sensors. In order to conduct complete intelligent automobile map data review, it is necessary to register and register the vehicle-mounted map data sensors (such as cameras, laser radars, millimeter wave radars, GNSS positioning, etc.).

[0063] Before data collection, serial communication type detection and adaptation, sensor data standard and protocol adaptation, and other matters need to be done. In the face of automobile geospatial data collection and application, the map review scheme based on edge computing first collects common serial communication types and data standard protocols for vehicle geospatial data sensors (such as cameras, lidar, millimeter wave radar, GNSS positioning, etc.), registers and registers them, and includes them in the data standard and protocol database. The steps of adaptive serial communication and data protocol adaptation include:

[0064] Step one, based on serial communication tools, adapt the communication serial port of the connected sensor to the network;

[0065] Step two, adapt the sensor data standard and protocol. This step requires the use of online access authorization information of the sensor itself, such as IP, User ID, password, etc. Online access authorization information will be obtained through configuration items, so that the database and adaptation calculation module are loosely coupled.

[0066] Step three, if the adaptation is authorized, the adaptation calculation process is ended; if the adaptation cannot be authorized, register and register the new protocol in the database, and return to step two.

[0067] Step S20: based on the preset point cloud target detection model, the laser point cloud data is detected, and the position corresponding to the first map data sensitive area in the laser point cloud data is determined;

[0068] After obtaining the geospatial data collected by each vehicle sensor of the intelligent vehicle, the laser point cloud data is detected based on the preset point cloud target detection model, and the position corresponding to the first map data sensitive area in the laser point cloud data is determined;

[0069] Further, the step of detecting the laser point cloud data based on the preset point cloud target detection model to determine the position corresponding to the first map data sensitive area in the laser point cloud data includes:

[0070] When the distance between the current position of the intelligent vehicle and the position of the predefined sensitive information is less than the first preset distance, the preset point cloud target detection model is called to detect the laser point cloud data, and the position corresponding to the first map data sensitive area in the laser point cloud data is determined. The first preset distance can be 200 meters, 250 meters, 300 meters, etc.

[0071] Specifically, for the sensitive area judgment calculation of lidar data, the judgment calculation process is as follows:

[0072] Detection_calc LiDAR(Location) = 1 (judging whether to need to perform point cloud target detection calculation based on the location of the intelligent vehicle);

[0073] s.t. {Distance(car, Location) < 200m (when the distance between the current location of the intelligent vehicle and the location of the sensitive information is less than 200 meters).

[0074] More specifically, the intelligent vehicle geospatial data review system can detect whether the distance between the current location of the intelligent vehicle and the location of the sensitive information is less than 200 meters, and when the distance between the current location of the intelligent vehicle and the location of the sensitive information is less than 200 meters, invoke a preset point cloud target detection model to perform point cloud target detection on the laser point cloud data to determine the position corresponding to the first map data sensitive region in the laser point cloud data.

[0075] In some embodiments, the point cloud target detection model is a deep learning model for processing three-dimensional point cloud data, which is commonly used to identify and locate interesting targets in point cloud data. Such models have important applications in the fields of autonomous driving, unmanned system autonomous navigation, industrial automation, etc. Point cloud is a data set composed of a large number of three-dimensional points, each point has its own coordinates (which can be polar coordinates or Cartesian coordinates) as well as color information (RGB values) and other attributes. Unlike image data, point cloud data is not a regular grid, but an unstructured data set, so it is more challenging to process. The point cloud target detection model includes SECOND, PointNet++ model, PointRCNN model, etc.

[0076] Step S30: performing target detection on the image data based on a preset image target detection model to determine the position corresponding to the second map data sensitive region in the image data;

[0077] After determining the position corresponding to the first map data sensitive region in the laser point cloud data, performing target detection on the image data based on a preset image target detection model to determine the position corresponding to the second map data sensitive region in the image data;

[0078] Further, the step of performing target detection on the image data based on a preset image target detection model to determine the position corresponding to the second map data sensitive region in the image data includes:

[0079] When it is detected that the distance between the current location of the intelligent vehicle and the location of the sensitive information is less than a second preset distance, or the set containing sensitive image information is not empty, or the set containing sensitive strings is not empty, target detection is performed on the image data based on a preset image target detection model to determine the location corresponding to the second map data sensitive area in the image data, wherein the second preset distance can be 150 meters.

[0080] Specifically, the calculation process for identifying sensitive areas in image data is as follows:

[0081] Detection_calc Camera (Location) = 1,

[0082] Among them, Detection_calc Camera (Location) = 1 indicates that the determination of whether to perform image target detection calculation is based on the location of the intelligent vehicle;

[0083]

[0084] Where Distance(car,Location)<150m means that the distance between the location of the intelligent vehicle and the location of the sensitive information is less than 150 meters.

[0085] "Set of sensitive picture is not empty" means that the set containing the sensitive picture target is not empty.

[0086] "Set of sensitive characters is not empty" means that the set containing sensitive characters is not empty.

[0087] More specifically, the intelligent vehicle geospatial data review system can detect whether the distance between the current location of the intelligent vehicle and the location of sensitive information is less than 150 meters, or whether the set containing sensitive image information is empty, or whether the set containing sensitive strings is empty. When it detects that the distance between the current location of the intelligent vehicle and the location of sensitive information is less than 150 meters, or the set containing sensitive image information is not empty, or the set containing sensitive strings is not empty, it performs target detection on the image data based on a preset image target detection model to determine the location corresponding to the sensitive area of ​​the second map data in the image data.

[0088] In some embodiments, the encryption processing of GNSS positioning data is performed in real time, therefore, there is no need for sensitive area judgment calculations.

[0089] Detection_calc gnss (Existence) = 1 (As long as the GNSS positioning result is valid, the positioning data encryption calculation is required)

[0090] In some embodiments, image object detection models are important technologies in the field of computer vision, designed to identify and locate different objects or targets in images. These models play a crucial role in many applications, such as security monitoring, autonomous driving, and medical image analysis. Image object detection mainly includes two main tasks: object classification: identifying different categories of objects present in an image, such as vehicles, pedestrians, and animals; and object localization: determining the position of the object in the image, typically using a bounding box to describe the object's position and size. Image object detection models include Mask R-CNN, YOLO, and others.

[0091] Step S40: Based on the location corresponding to the first map data sensitive area in the laser point cloud data and the location corresponding to the second map data sensitive area in the image data, determine the set of sensitive information to be reviewed in the geospatial data, and determine the sensitive information in the geospatial data according to the set of sensitive information to be reviewed;

[0092] After determining the location corresponding to the first sensitive area in the laser point cloud data and the location corresponding to the second sensitive area in the image data, based on the location corresponding to the first sensitive area in the laser point cloud data and the location corresponding to the second sensitive area in the image data, a set of sensitive information to be reviewed in the geospatial data is determined, and the sensitive information in the geospatial data is determined according to the set of sensitive information to be reviewed.

[0093] Furthermore, based on the locations corresponding to the first sensitive area in the laser point cloud data and the locations corresponding to the second sensitive area in the image data, a set of sensitive information to be reviewed in the geospatial data is determined. The step of determining the sensitive information in the geospatial data according to the set of sensitive information to be reviewed includes:

[0094] Step S401: Based on the location corresponding to the first map data sensitive area in the laser point cloud data and the location corresponding to the second map data sensitive area in the image data, determine the sensitive information that needs to be focused on in the current stage in the sensitive geographic information database, so as to determine the set M of sensitive information to be reviewed;

[0095] Step S402: Read a new information record from the set of sensitive information to be reviewed M, and denot it as sensitive information record a to be reviewed;

[0096] Step S403: Perform image retrieval analysis on the geospatial data of the current frame, calculate and match it with the sensitive information record a to be reviewed, and determine whether the geospatial data of the current frame contains sensitive information associated with the sensitive information record a to be reviewed.

[0097] Step S404: If the geospatial data of the current frame contains a sensitive information record associated with the sensitive information record a to be reviewed, determine the type of the sensitive information record a to be reviewed and its coordinate position in the geospatial data of the current frame, and add it to the sensitive information detection result set N.

[0098] Specifically, please refer to Figure 4 Based on the locations corresponding to the first sensitive area in the laser point cloud data and the locations corresponding to the second sensitive area in the image data, the calculation process for detecting sensitive geographic information is as follows:

[0099] Step 1: Acquire geospatial data (raw sensor data) from vehicle sensors;

[0100] Step 2: Determine whether the geospatial data (raw sensor data) at the current frame rate requires sensitive geographic information detection. If yes, proceed to Step 3; otherwise, proceed to Step 7.

[0101] Step 3: Based on the combination of time, spatial information, and other conditions, and according to the corresponding locations of the first map data sensitive areas in the laser point cloud data and the second map data sensitive areas in the image data, a query is performed in the sensitive geographic information database to obtain the sensitive information that needs attention at the current stage. The sensitive information includes sensitive images and sensitive text, etc., to determine the set M of sensitive information to be reviewed. The sensitive geographic information database is a pre-designed and constructed information infrastructure, which is generally built and maintained by the competent authority and provides external access services.

[0102] Step 4: Read a new information record from the set of sensitive information to be reviewed M, and denot it as sensitive information record a to be reviewed, until all records in set M have been traversed;

[0103] Step 5: Perform image retrieval analysis on the geospatial data of the current frame and calculate and match it with the sensitive information record a to be reviewed. Determine whether the geospatial data of the current frame contains sensitive information associated with the sensitive information record a to be reviewed. If it does, proceed to Step 6; otherwise, return to Step 4.

[0104] Step 6: Record the information type of the sensitive information record a to be reviewed and its coordinate position (Bounding BOX) in the current frame data, and add it to the sensitive information detection result set N.

[0105] Step 7: Summarize the sensitive information detection results into set N and use it as the output. End.

[0106] Step S50: Filter the sensitive information in the geospatial data according to the preset sensitive information filtering algorithm to complete the review of the intelligent vehicle map data.

[0107] After determining the sensitive information in the geospatial data based on the set of sensitive information to be reviewed, the sensitive information in the geospatial data is filtered according to a preset sensitive information filtering algorithm to complete the review of the intelligent vehicle map data.

[0108] Furthermore, the process of filtering sensitive information in the geospatial data according to a preset sensitive information filtering algorithm to complete the review of intelligent vehicle map data includes:

[0109] Step S501: Obtain sensitive information detection result records from the sensitive information detection result set N, wherein the sensitive information detection result records include sensitive information type and sensitive information region;

[0110] Step S502: Determine the preset expansion radius, and expand the circle outward from any point O(i,j) in the sensitive information area according to the preset expansion radius to form a circle;

[0111] Step S503: Remove the intersection area between the sensitive information area and the circle to filter the sensitive information in the geospatial data.

[0112] Specifically, please refer to Figure 5 Image sensitive information region filtering refers to the intelligent computation process of removing sensitive information from the current frame data containing sensitive information. Specific steps include:

[0113] Step 1: Read the sensitive information detection result set N, and perform structured processing on the data in the sensitive information detection result set N;

[0114] Step 2: Obtain a new record from the sensitive information detection result set N, including the data type of the sensitive information and the sensitive information region (Bbox_n). Record the four corner coordinates of the sensitive information region (Bbox_n), denoted as Max_x_n, Min_x_n, Max_y_n, and Min_y_n. Max_x_n represents the maximum X-coordinate of the outer frame of the nth record, Min_x_n represents the minimum X-coordinate of the outer frame of the nth record, Max_y_n represents the maximum Y-coordinate of the outer frame of the nth record, and Min_y_n represents the minimum Y-coordinate of the outer frame of the nth record.

[0115] Step 3: Please refer to Figure 6 Take any point within the sensitive information region (Bbox_n) and statistically obtain its RGB value. The calculation method is as follows:

[0116]

[0117]

[0118] Among them, RGB_value i,j The RGB value of a pixel with pixel coordinates O(i,j) after calculation. This represents the summation and average of pixel values ​​within the sensitive information region; other symbols have the following meanings:

[0119] For any point O(i,j) within the sensitive information region (Bbox), expand outwards to the surrounding area. Remove the intersection area between the Bbox and the circle, as shown in the figure, which is the black dot area. If the number of pixels in the black dot area is greater than 100, or the number of pixels in the black dot area accounts for more than 5% of the total number of pixels in the image, it can be determined that the RGB value of the pixels in the black dot area is strongly correlated with the RGB value of O(i,j).

[0120] The regions with more than 100 pixels in the black dot region, or regions where the number of pixels in the black dot region accounts for more than 5% of the total number of pixels in the image, are removed to filter sensitive information regions in the image data.

[0121] Step 4: Return to Step 2 until all records in the sensitive information detection result set N have been processed, and then end the calculation.

[0122] In some embodiments, the sensitive information includes one or more of the following: sensitive geographical indication images, sensitive administrative planning images, sensitive geographical indication logos, and sensitive administrative planning logos.

[0123] As can be seen from the above embodiments, compared with the prior art, this application addresses the map review of finished autonomous driving map data in the prior art. However, the review often focuses on the post-processing results of intelligent vehicle map data, and lacks supervision over the data collection, transmission, and preprocessing processes of intelligent vehicle map data sensors. This makes it impossible to achieve full-process review and monitoring of intelligent vehicle geospatial data. This application has the following beneficial effects, including but not limited to:

[0124] Firstly, the intelligent vehicle map data review method proposed in this application, based on the feature analysis of intelligent vehicle geospatial data sensors and data processing, and the objectives of map review, proposes a real-time intelligent map data review method based on edge intelligent computing. Through embedded computing, it adaptively adapts to the intelligent vehicle's map data sensors to monitor the intelligent vehicle's geospatial data acquisition process. Using image intelligent algorithms, such as image target detection and tracking, and data mining, it retrieves sensitive targets in the image and repairs the image. Through data classification and feature extraction from the sensitive target information database, it achieves high-accuracy, low-time-consumption extraction of sensitive information, as well as the identification of omissions and inconsistencies in the map data.

[0125] Secondly, the intelligent vehicle map data review method of this application is based on the autonomous driving map data review system architecture for real-time review. It loosely couples intelligent vehicle geospatial data collection, sensitive map element identification, and data repair, thereby achieving convenient and efficient map review work management and improving the executability and interpretability of the map review process.

[0126] Third, the intelligent vehicle map data review method of this application is based on data classification and feature extraction of a sensitive target information database, thereby reducing the requirements for hardware platform storage and computing performance of intelligent computing for map review, while responding more conveniently to the requirements of map review policies and regulatory agencies.

[0127] Please see Figure 7A smart car map data verification device provided for one of the purposes of this application includes a sensor data acquisition module 1100, a first sensitive area determination module 1200, a second sensitive area determination module 1300, a sensitive information determination module 1400, and a map data verification module 1500. The system includes the following modules: a sensor data acquisition module 1100, configured to acquire geospatial data collected by various onboard sensors of the intelligent vehicle in response to an intelligent vehicle map data review instruction; a first sensitive area determination module 1200, configured to perform point cloud target detection on the laser point cloud data based on a preset point cloud target detection model to determine the location corresponding to the first map data sensitive area in the laser point cloud data; a second sensitive area determination module 1300, configured to perform target detection on the image data based on a preset image target detection model to determine the location corresponding to the second map data sensitive area in the image data; a sensitive information determination module 1400, configured to determine the set of sensitive information to be reviewed in the geospatial data based on the locations corresponding to the first map data sensitive area in the laser point cloud data and the locations corresponding to the second map data sensitive area in the image data, and to determine the sensitive information in the geospatial data according to the set of sensitive information to be reviewed; and a map data review module 1500, configured to filter the sensitive information in the geospatial data according to a preset sensitive information filtering algorithm to complete the review of the intelligent vehicle map data.

[0128] Based on any embodiment of this application, please refer to Figure 8 Another embodiment of this application provides an electronic device, which can be implemented by a computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected via a system bus. The computer-readable storage medium stores an operating system, a database, and computer-readable instructions. The database may store a sequence of control information. When the computer-readable instructions are executed by the processor, the processor can implement a method for verifying intelligent vehicle map data. The processor of the computer device provides computing and control capabilities to support the operation of the entire computer device. The memory of the computer device may store computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor can execute the intelligent vehicle map data verification method of this application. The network interface of the computer device is used for communication with a terminal. Those skilled in the art will understand that… Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0129] In this embodiment, the processor is used to execute... Figure 7 The system defines the specific functions of each module and its sub-modules. The memory stores the program code and various data required to execute these modules or sub-modules. The network interface is used for data transmission between the user terminal and the server. In this embodiment, the memory stores the program code and data required to execute all modules / sub-modules in the intelligent vehicle map data review device of this application. The server can call the server's program code and data to execute the functions of all sub-modules.

[0130] This application also provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the intelligent vehicle map data review method described in any embodiment of this application.

[0131] This application also provides a computer program product, including a computer program / instructions that, when executed by one or more processors, implement the steps of the intelligent vehicle map data review method described in any embodiment of this application.

[0132] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium can be a computer-readable storage medium such as a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM).

[0133] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for reviewing intelligent vehicle map data, characterized in that, Includes the following steps: In response to the intelligent vehicle map data review instruction, the system acquires geospatial data collected by various on-board sensors of the intelligent vehicle, wherein the geospatial data includes at least image data and laser point cloud data. Point cloud target detection is performed on the laser point cloud data to determine the location corresponding to the first map data sensitive area in the laser point cloud data; Target detection is performed on the image data to determine the location corresponding to the sensitive area of ​​the second map data in the image data; Based on the location corresponding to the first map data sensitive area in the laser point cloud data and the location corresponding to the second map data sensitive area in the image data, a set of sensitive information to be reviewed in the geospatial data is determined, and sensitive information in the geospatial data is determined according to the set of sensitive information to be reviewed. Sensitive information in the geospatial data is filtered out to complete the review of the intelligent vehicle map data.

2. The intelligent vehicle map data verification method according to claim 1, characterized in that, Before the steps of acquiring geospatial data collected by various onboard sensors of a smart car, the process also includes: The system detects whether the geospatial data collected by each vehicle-mounted sensor conforms to the preset serial communication type and sensor data standard. If it does, the geospatial data collected by each vehicle-mounted sensor is input into the intelligent vehicle geospatial data review system for processing.

3. The intelligent vehicle map data verification method according to claim 1, characterized in that, The steps of performing point cloud target detection on the laser point cloud data and determining the location corresponding to the first map data sensitive area in the laser point cloud data include: When the distance between the current location of the intelligent vehicle and the location of the sensitive information is less than a first preset distance, a preset point cloud target detection model is invoked to perform point cloud target detection on the laser point cloud data, and the location corresponding to the first map data sensitive area in the laser point cloud data is determined.

4. The intelligent vehicle map data verification method according to claim 1, characterized in that, The step of performing target detection on the image data and determining the location corresponding to the sensitive area of ​​the second map data in the image data includes: When it is detected that the distance between the current location of the intelligent vehicle and the location of the sensitive information is less than a second preset distance, or the set containing sensitive image information is not empty, or the set containing sensitive strings is not empty, target detection is performed on the image data based on a preset image target detection model to determine the location corresponding to the sensitive area of ​​the second map data in the image data.

5. The intelligent vehicle map data verification method according to claim 1, characterized in that, Based on the locations corresponding to the first sensitive area in the laser point cloud data and the locations corresponding to the second sensitive area in the image data, a set of sensitive information to be reviewed in the geospatial data is determined. The step of determining the sensitive information in the geospatial data according to the set of sensitive information to be reviewed includes: Based on the location corresponding to the first map data sensitive area in the laser point cloud data and the location corresponding to the second map data sensitive area in the image data, the sensitive information that needs to be focused on at the current stage is determined in the sensitive geographic information database to determine the set M of sensitive information to be reviewed; Read a new information record from the set of sensitive information to be reviewed M, and denote it as sensitive information record a to be reviewed; The geospatial data of the current frame is subjected to image retrieval analysis and matched with the sensitive information record a to be reviewed to determine whether the geospatial data of the current frame contains sensitive information associated with the sensitive information record a to be reviewed. If the geospatial data of the current frame contains a sensitive information record associated with the sensitive information record a to be reviewed, determine the type of the sensitive information record a to be reviewed and its coordinate position in the geospatial data of the current frame, and add it to the sensitive information detection result set N.

6. The intelligent vehicle map data verification method according to claim 5, characterized in that, The steps for filtering sensitive information in the geospatial data to complete the review of intelligent vehicle map data include: Obtain sensitive information detection result records from the sensitive information detection result set N, wherein the sensitive information detection result records include sensitive information type and sensitive information region; A preset expansion radius is determined, and any point O(i,j) within the sensitive information area is used as the center to expand outwards to form a circle according to the preset expansion radius; The intersection area between the sensitive information area and the circle is removed to filter out the sensitive information in the geospatial data.

7. The intelligent vehicle map data verification method according to any one of claims 1 to 6, characterized in that, The sensitive information includes one or more of the following: sensitive geographical indication images, sensitive administrative planning images, sensitive geographical indication logos, and sensitive administrative planning logos.

8. A smart car map data verification device, characterized in that, include: The sensor data acquisition module is configured to respond to the intelligent vehicle map data review instruction and acquire geospatial data collected by various on-board sensors of the intelligent vehicle, wherein the geospatial data includes at least image data and laser point cloud data. The first sensitive area determination module is configured to perform point cloud target detection on the laser point cloud data based on a preset point cloud target detection model, and determine the location corresponding to the first map data sensitive area in the laser point cloud data. The second sensitive area determination module is configured to perform target detection on the image data based on a preset image target detection model, and determine the location corresponding to the second map data sensitive area in the image data. The sensitive information determination module is configured to determine the set of sensitive information to be reviewed in the geospatial data based on the location corresponding to the first map data sensitive area in the laser point cloud data and the location corresponding to the second map data sensitive area in the image data, and to determine the sensitive information in the geospatial data according to the set of sensitive information to be reviewed. The map data review module is configured to filter sensitive information in the geospatial data according to a preset sensitive information filtering algorithm in order to complete the review of the intelligent vehicle map data.

9. An electronic device comprising a central processing unit and a memory, characterized in that, The central processing unit is used to invoke and run a computer program stored in the memory to perform the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores, in the form of computer-readable instructions, a computer program implemented according to any one of claims 1 to 7, which, when invoked by a computer, executes the steps included in the corresponding method.

Images