Data processing method and device, storage medium and electronic equipment

Abstract

The embodiment of the invention provides a data processing method and device, a storage medium and electronic equipment, and the method comprises the steps: obtaining the operation data of a hovercar under the condition that the hovercar is running, the hovercar communicates with a cloud end, and the hovercar is provided with a pre-trained sensitive recognition model; performing sensitive identification processing on the operation data through a sensitive identification model to obtain a sensitive identification result; under the condition that the sensitive identification result indicates that the sensitive data exists in the operation data, desensitizing the sensitive data to obtain desensitized data; and encrypting other data except the sensitive data in the desensitized data and the operation data, and sending a target data packet obtained by encryption to the cloud. According to the data processing method and device, the problem that the leakage risk is high when original operation data is transmitted in the prior art is solved, and then the effect of improving the data processing safety is achieved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data processing technology, and more specifically, to a data processing method and apparatus, a storage medium, and an electronic device. Background Art

[0002] With the development of intelligent transportation systems and low-altitude economy, flying cars, as an emerging means of transportation, collect a large amount of data during their operation, including but not limited to operating location, environmental perception images and videos, point cloud data, etc., which makes it possible to achieve efficient and safe autonomous driving.

[0003] In the related art, data processing in the flying car field primarily occurs in the cloud, which communicates with the flying car. The cloud directly deletes specific types of data from the flying car's raw operational data. However, this method, which directly uploads unprocessed raw data to the cloud, is susceptible to hacker attacks or illegal interception during transmission, especially in wireless network environments. Therefore, this related art poses a high risk of leakage during the transmission of raw operational data. Summary of the Invention

[0004] The embodiments of the present application provide a data processing method and device, a storage medium, and an electronic device to at least solve the technical problem in the related art that the risk of leakage during transmission of original running data is relatively high.

[0005] According to one aspect of an embodiment of the present application, a data processing method is provided, including:

[0006] When a flying car is running, operating data of the flying car is obtained, wherein the flying car communicates with a cloud and a pre-trained sensitive recognition model is deployed on the flying car; sensitive recognition processing is performed on the operating data using the sensitive recognition model to obtain a sensitive recognition result; when the sensitive recognition result indicates that sensitive data exists in the operating data, the sensitive data is desensitized to obtain desensitized data; the desensitized data and other data in the operating data except the sensitive data are encrypted, and the encrypted target data packet is sent to the cloud.

[0007] According to another aspect of an embodiment of the present application, a data processing device is also provided, including: an acquisition unit, for acquiring operating data of a flying car when the flying car is running, wherein the flying car communicates with the cloud and is deployed with a pre-trained sensitive recognition model; an identification unit, for performing sensitive recognition processing on the operating data through the sensitive recognition model to obtain a sensitive recognition result; a first desensitizing unit, for performing desensitizing processing on the sensitive data to obtain desensitized data when the sensitive recognition result indicates that sensitive data exists in the operating data; an encryption unit, for encrypting the desensitized data and other data in the operating data except the sensitive data, and sending the encrypted target data packet to the cloud.

[0008] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when run.

[0009] According to another aspect of the embodiments of the present application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps of any of the above-described method embodiments.

[0010] According to another aspect of the embodiments of the present application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the steps of any of the above method embodiments through the computer program.

[0011] Through the present application and the embodiments provided by the present application, when a flying car is running, a pre-trained sensitive identification model deployed on the flying car is used to perform sensitive identification on the operating data, and when the sensitive identification result indicates the presence of sensitive data, the sensitive data is desensitized, and then the desensitized data and other data in the operating data except the sensitive data are encrypted to transmit the encrypted target data packet to the cloud, thereby reducing the risk of sensitive information leakage before the data is transmitted to the cloud, and to a certain extent solving the problem of relatively high leakage risk during transmission of original operating data in related technologies, improving the security of data processing, and also reducing the data processing burden on the cloud. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a schematic diagram of an application scenario of a data processing method according to an embodiment of the present application;

[0013] Figure 2 is a flowchart of an optional data processing method according to an embodiment of the present application;

[0014] Figure 3 is a schematic diagram of an optional data processing method according to an embodiment of the present application;

[0015] Figure 4 is a flowchart of another optional data processing method according to an embodiment of the present application;

[0016] Figure 5 is a structural block diagram of an optional data processing device according to an embodiment of the present application;

[0017] Figure 6 This is a block diagram of a computer system structure of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0019] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0020] According to one aspect of the embodiment of the present application, a data processing method is provided. Optionally, in this embodiment, the above data processing method can be applied to, but is not limited to, Figure 1The hardware environment shown includes a terminal device 102 and a server 104. The server 104 can be connected to the terminal device 102 via a network and can be used to provide services (e.g., application services, etc.) for the terminal device 102 or a client installed on the terminal device 102. A database can be set on the server 104 or independently of the server 104 to provide data storage services for the server 104.

[0021] The aforementioned network may include, but is not limited to, at least one of the following: a wired network and a wireless network. The aforementioned wired network may include, but is not limited to, at least one of the following: a wide area network, a metropolitan area network, or a local area network. The aforementioned wireless network may include, but is not limited to, at least one of the following: Wi-Fi (Wireless Fidelity) and Bluetooth. The terminal device 102 may be, but is not limited to, a PC (Personal Computer), a mobile phone, a tablet computer, etc. The server 104 may be, but is not limited to, a cloud server, a server cluster, or other server types.

[0022] The data processing method of the embodiment of the present application can be jointly executed by the server 104 and the terminal device 102. Specifically, it can be jointly executed by the flying car (i.e., the vehicle side) and the cloud.

[0023] Figure 2 is a flow chart of an optional data processing method according to an embodiment of the present application, such as Figure 2 As shown, the process of the method may include the following steps:

[0024] Step S202: Acquire operating data of the flying car while the flying car is in operation, wherein the flying car communicates with the cloud and a pre-trained sensitive recognition model is deployed on the flying car;

[0025] Step S204, performing sensitivity identification processing on the operating data using a sensitivity identification model to obtain a sensitivity identification result;

[0026] Step S206: If the sensitivity identification result indicates that sensitive data exists in the operating data, desensitizing the sensitive data to obtain desensitized data;

[0027] In step S208, the desensitized data and the other data in the running data except the sensitive data are encrypted, and the encrypted target data packet is sent to the cloud.

[0028] The data processing method in this embodiment can be applied to the field of data processing technology, and applied to the scenarios of real-time operation monitoring of flying cars and intelligent driving data security management.

[0029] In the related art, data processing in the flying car field primarily occurs in the cloud, which communicates with the flying car. The cloud directly deletes specific types of data transmitted by the flying car. However, this method directly uploads raw, unprocessed operating data to the cloud, making it vulnerable to hacker attacks or illegal interception during transmission, especially in wireless network environments. Therefore, this related art presents a high risk of leakage during the transmission of raw operating data.

[0030] In order to at least partially solve the above technical problems, in this embodiment, when the flying car is running, the operating data is sensitively identified through a pre-trained sensitive identification model deployed on the flying car, and when the sensitive identification result indicates the presence of sensitive data, the sensitive data is desensitized, and then the desensitized data and other data in the operating data except the sensitive data are encrypted to transmit the encrypted target data packet to the cloud, thereby reducing the risk of sensitive information leakage before the data is transmitted to the cloud, solving to a certain extent the problem of relatively high leakage risk during the transmission of original operating data, improving the security of data processing, and also reducing the data processing burden on the cloud.

[0031] It should be noted that a flying car can be a vehicle capable of both land and air travel, combining the functions of a car and an aircraft with a highly automated driving system. Flying cars communicate with remote servers (i.e., the cloud) via wireless network technology to exchange data, upload operational data, and receive instructions.

[0032] The flying car is equipped with multiple different sensor devices to collect operating data in real time and establish a connection with the cloud through wireless communication technologies such as 4G / 5G.

[0033] Of course, the flying car also has a pre-trained sensitive identification model deployed. This pre-trained sensitive identification model is a deep learning-based neural network model used to identify sensitive information that may be present in operational data, such as personal privacy data and location data of specific geographically sensitive areas. Optionally, there can be multiple sensitive identification models. In this case, one sensitive identification model is used to identify at least one type of sensitive data, and different sensitive identification models are used to identify different types of sensitive data.

[0034] Optionally, the sensitive identification model can be trained using transfer learning. First, a deep neural network model, such as ResNet, EfficientNet, or another convolutional neural network (CNN), is pre-trained on a large-scale, general dataset to acquire rich feature representation capabilities. This pre-trained model is then fine-tuned and trained using sensitive datasets specific to flying car operations, such as images and point cloud data containing sensitive information such as sensitive areas and personal sensitive data. By annotating sensitive areas in the data, the model learns to identify these sensitive features in new, unseen data, improving its generalization capabilities. Data augmentation techniques, such as rotation, flipping, and scaling, can be employed during training to increase model robustness. The sensitive identification model can be a complex deep learning model, such as one based on the Transformer architecture, incorporating attention mechanisms and multimodal data processing capabilities to more effectively understand and identify sensitive information in flying car operation data, improving recognition accuracy and processing speed.

[0035] Optionally, when sensitive data is identified as existing in the operating data, different desensitizing strategies can be adopted according to the type of sensitive data to desensitize the sensitive data to obtain desensitized data corresponding to the sensitive data, wherein the desensitized data can be the desensitized data obtained by desensitizing the sensitive data in the operating data.

[0036] Optionally, after obtaining the desensitized data, the desensitized data and other data in the running data except sensitive data can be encrypted using a preset encryption algorithm to obtain an encrypted target data packet. The encrypted target data packet includes the target data, and the target data includes the desensitized data and other data in the running data except sensitive data.

[0037] Optionally, desensitized data and non-sensitive data are encrypted using an encryption algorithm (such as the national encryption standard SM4) to generate ciphertext data, the encrypted target data is packaged and processed to obtain an encrypted target data packet, and sent to the cloud through a secure communication protocol to ensure the confidentiality and integrity of the data during transmission.

[0038] Optionally, to further ensure data transmission security, a dual encryption technology combining national encryption standards (such as SM2 / SM3 / SM4) with private encryption protocols can be employed during secure communication between the flying car and the cloud to ensure confidentiality, integrity, and authenticity of data transmission. Specifically, before the flying car transmits the encrypted target data packet to the cloud, mutual identity authentication can be performed to ensure the legitimacy of the communication. The cloud then decrypts the data packet using the same private decryption algorithm and further verifies the data's compliance. Furthermore, to ensure data security within the cloud, intra-cloud data transmission utilizes VPC peering technology. This technology, through dedicated lines and IPSec VPN technology, encrypts all backhaul links, enabling private and secure data flow between different cloud services. A key management system (KMS) operates independently, automatically allocating and regularly updating encryption and decryption keys, preventing security vulnerabilities caused by human intervention and ensuring the complete security and controllability of communication between the flying car and the cloud. Through these measures, secure communication between the cloud and the vehicle is achieved, effectively preventing data leakage, tampering and unauthorized access during transmission, and ensuring the compliance and security of flying car operating data.

[0039] Through the embodiments provided by the present application, when a flying car is running, a pre-trained sensitive identification model deployed on the flying car is used to perform sensitive identification on the operating data, and when the sensitive identification result indicates the presence of sensitive data, the sensitive data is desensitized, and then the desensitized data and other data in the operating data except the sensitive data are encrypted to transmit the encrypted target data packet to the cloud, thereby reducing the risk of sensitive information leakage before the data is transmitted to the cloud, and to a certain extent solving the problem of relatively high leakage risk during transmission of original operating data in related technologies, improving the security of data processing, and also reducing the data processing burden on the cloud.

[0040] In an exemplary embodiment, there are multiple sensitive recognition models, and the multiple sensitive recognition models include a sensitive area recognition model and an image recognition model. The flying car is provided with at least one visual perception device, and the operation data includes at least one visual perception device for collecting operation images of the flying car during operation; step S204 includes: inputting the operation data into the sensitive area recognition model, and outputting a predicted recognition result, wherein the predicted recognition result is used to indicate whether sensitive area data exists in the operation data; inputting the operation image into the image recognition model, and outputting an image recognition result corresponding to the operation image, wherein the image recognition result is used to indicate whether personal sensitive data exists in the operation image, and the personal sensitive data includes at least one of the following: facial information, license plate information.

[0041] It should be noted that there are multiple sensitive identification models, including sensitive area identification models and image recognition models, which can be used to identify sensitive area data and personal sensitive data collected by visual sensors, respectively. Sensitive area data can primarily be data related to specific areas that require a certain degree of confidentiality, such as location data.

[0042] Pre-trained Sensitive Area Identification Model The sensitive area identification model can be mainly used to identify relevant data of some specific sensitive areas. Sensitive area data can involve relevant data of units and facilities related to public safety, and relevant data of some special facilities related to traffic safety. Among them, relevant data of some special facilities related to traffic safety can include relevant data such as undisclosed harbors, ports, airports, and other related units, and of course, it can also include precise data such as public airports, waterway depths, lock dimensions, reservoir capacity, bridge-related data (such as height limits, width limits, clearance, load capacity and slope attributes), tunnel-related data (such as tunnel height and width attribute data), highway-related data (such as road paving material attribute data), and reservoir-related attribute data (such as storage capacity attributes and height data).

[0043] Personal sensitive data can be personal privacy data, such as facial information and license plate information. Through image recognition models, it can be ensured to a certain extent that flying cars will not leak personal privacy when collecting and transmitting image data.

[0044] Optionally, the flying car is equipped with multiple sensing devices, such as cameras to capture visual information, LiDAR to build three-dimensional point cloud maps, millimeter-wave radar and ultrasonic sensors to detect obstacles, infrared cameras to enhance night vision, GPS / Beidou positioning systems and IMU to ensure precise navigation and attitude control, and environmental perception sensors to collect meteorological data, all of which jointly support its intelligent and safe flight.

[0045] Among them, visual perception devices can be installed on flying cars to capture images and videos of the surrounding environment, provide real-time visual information for flying cars, and support functions such as intelligent driving and environmental perception.

[0046] Optionally, the operational data may be multimodal data, including images, point clouds, text, and other modal data. The operational data may include a collection of information collected and generated by the flying car during operation. It may be multimodal data and specifically include the following types: image data collected by visual sensors, point cloud data collected by lidar equipment, flight parameters of the flying car, such as altitude, speed, acceleration, attitude angles (pitch, yaw, roll), etc., as well as text data including operational logs and system status.

[0047] The operating image can be image data collected by the flying car's visual perception equipment during the flight, which is used to assist in autonomous driving, environmental monitoring and subsequent data analysis.

[0048] The prediction and identification result may be a prediction result identified by a sensitive area identification model, wherein the prediction and identification result may be used to indicate whether the operation data includes sensitive area data.

[0049] The image recognition result may be the result output after the running image is processed by the image recognition model, wherein the image recognition result is used to indicate whether personal sensitive data is stored in the running image.

[0050] Optionally, the sensitive area recognition model can be a multimodal data recognition model. The training process of the sensitive area recognition model can include: a large amount of geospatial data, including clearly labeled non-sensitive and sensitive areas, multimodal data such as images, point clouds, and text, as well as the specific coordinates and description information of sensitive areas. The acquired multimodal data is trained and verified based on the Transformer architecture or similar deep learning models. The pre-trained model can also be fine-tuned on the sensitive area dataset, and the model parameters can be updated using the backpropagation algorithm to enhance the sensitive area recognition model's ability to recognize sensitive area data.

[0051] Optionally, the training process of the image recognition model may include: obtaining an image dataset of personal sensitive information, where the image dataset includes different types of personal sensitive data, and the image dataset needs to cover images in different environments (different lighting, different angles, and different occlusion conditions) to improve the generalization ability of the model. Data preprocessing is performed on the images in the image dataset, such as size standardization, grayscale conversion, brightness and contrast adjustment, etc., to enhance the model training effect. A high-performance convolutional neural network (CNN) is used for pre-training and verification using a public image dataset (such as ImageNet) to learn common image features to obtain a pre-trained image recognition model.

[0052] Through the above training process, the two models are able to acquire sufficient capabilities to accurately identify personal sensitive information in sensitive areas and images in real time during the actual operation of flying cars, thereby performing corresponding compliance processing to ensure the security and compliance of flying car data.

[0053] This embodiment deploys a sensitive area recognition model and an image recognition model on the flying car. The sensitive area recognition model can accurately locate specific geographic coordinate data, while the image recognition model can effectively identify facial information and license plate information. This allows for rapid desensitization or deletion of sensitive data on the vehicle itself. Sensitive data is processed on the flying car itself, and the encrypted data packets are more securely transmitted to the cloud, reducing the risk of data interception and decryption during transmission. Furthermore, by offloading some data processing tasks to the local model on the flying car itself, the amount of data uploaded to the cloud is reduced, thereby alleviating the computing pressure on the cloud server.

[0054] In an exemplary embodiment, step S206 includes: when the sensitive data is sensitive area data, performing a first desensitization process on the sensitive area data to obtain first desensitized data, wherein the first desensitization process includes at least one of the following: deletion process, replacement process; when the sensitive data is personal sensitive data, performing a second desensitization process on the personal sensitive data to obtain second desensitized data, wherein the second desensitization process includes at least one of the following: blocking process, blurring process, and replacement process.

[0055] It should be noted that sensitive areas may include sensitive area data and personal sensitive data. Sensitive area data may mainly be relevant data of some specific areas that need to be kept confidential to a certain extent. The first desensitization process can be used to delete or replace this data. The first desensitization process may include at least one of the following: deletion process, replacement process. Among them, the deletion process is to directly remove the geographic coordinates, images or point cloud data of the sensitive area to ensure that the operating data does not contain any specific information of the sensitive area. The replacement process is to replace the data of the sensitive area with non-sensitive, general or fictitious data, such as using virtual coordinates to replace real coordinates, to avoid the exposure of direct information, but keep the data structure intact for subsequent analysis and use.

[0056] The second desensitization process may include at least one of the following: occlusion processing, blurring processing, and replacement processing. Among them, occlusion processing refers to using a specific pattern or color to cover the sensitive information in the location area where the sensitive information is located in the running image, such as blocking the face or license plate to prevent personal information from being identified. Blurring processing can be blurring the location area where the sensitive information is located in the running image (such as Gaussian blurring) to make it visually difficult to identify, but still retain the details of the surrounding environment. Replacement processing can be replacing sensitive information, such as facial features, license plate numbers, in the running image with randomly generated or predefined images (for example, replacement of mosaic images).

[0057] Specifically, if Figure 3 As shown, the process of desensitizing personal sensitive information can be referred to Figure 3 , taking occlusion processing as an example of desensitization method for explanation, wherein the first image is an operation image collected during the operation of the flying car, and the first image includes facial information and vehicle license plate information. When it is recognized that the first image contains facial information and vehicle license plate information, the area where the face is located is occluded by a preset image (i.e., the gray part in the figure), and the license plate information is occluded by the preset image.

[0058] In one example, in the image data of a flying car, the AI model is used to accurately locate the position of a face or license plate, and occlusion technology is applied, such as using a pattern to cover the face area and filling the license plate position with a specific color, to ensure that personal data is unrecognizable in the image without affecting the flying car's environmental perception capabilities.

[0059] In another example, image processing technology is used to perform Gaussian blur processing on the license plate area or face, making the license plate number blurred but still maintaining the clarity of the surrounding environment, thereby protecting personal information while retaining the analytical value of the image.

[0060] Through this embodiment, the deletion and replacement of sensitive area data, as well as the blocking, blurring and replacement of personal sensitive data, reduce the exposure of sensitive information from the source.

[0061] In an exemplary embodiment, the operating data includes operating position data; the above method also includes: performing coordinate deflection processing on the operating position data to obtain virtual coordinate data corresponding to the operating position data, wherein the coordinate deflection processing includes at least one of the following: coordinate conversion processing, coordinate mapping processing.

[0062] It should be noted that the operation data may include operation position data, which may be represented by GPS or Beidou coordinate information of the flying car. The virtual coordinate data may be coordinate data corresponding to the operation position after coordinate deflection.

[0063] Coordinate deflection processing can include coordinate conversion processing and coordinate mapping processing. Coordinate conversion processing can be to transform the position information in the original coordinate system into another unrelated or fictitious coordinate system through data changes, ensuring that the outside world cannot directly trace back to the actual position from the converted coordinate data. It can also include the application of coordinate system conversion and encryption algorithms.

[0064] Coordinate mapping processing can map the original coordinates to a regular virtual coordinate that has a certain deviation from the actual location based on pre-set mapping rules. The design of the mapping rules needs to consider the availability and security of the data to ensure that the data can still be used in specific business scenarios without leaking sensitive location information.

[0065] In one example, after the operation data is obtained, the original operation position data can be encrypted according to a certain offset and direction using a specific encryption algorithm and converted into virtual coordinate data.

[0066] In another example, after acquiring the operational data, the original coordinates can be mapped to virtual coordinates that are offset from the actual location according to a preset mapping rule. For example, for a specific sensitive area, a fixed offset and direction can be set, and all operational location data in that area can be mapped to the virtual coordinate system to mask the actual location.

[0067] Optionally, the coordinate deflection processing of the operating position data can be performed before the sensitive identification model is used to perform sensitive identification processing on the operating data of the flying car, or after the sensitive identification model is used to perform sensitive identification processing on the operating data of the flying car, as long as it is before being sent to the cloud. This application does not limit this.

[0068] Optionally, before using the sensitive recognition model to perform sensitive recognition processing on the flying car's operating data and obtaining the sensitive recognition result, the operating position data is subjected to coordinate deflection processing.

[0069] Through this embodiment, through coordinate deflection processing, the flying car can ensure that sensitive information of the operating position data is not directly obtained. Even if the data is intercepted, the virtual coordinate data cannot be directly corresponded to the real position, which significantly improves the security of the data.

[0070] In an exemplary embodiment, the operating data includes operating position data and operating status data. The flying car is deployed with a designated airspace database, and the designated airspace database records a set of designated airspaces. The above method also includes: when the operating status data indicates that the flying car is in a flying state, identifying whether the flying car is in a designated airspace in the designated airspace database based on the operating position data; when the flying car is in the designated airspace, deleting the target operating data, wherein the target operating data is the operating data of the flying car in the designated airspace.

[0071] It should be noted that operational data may include operational location data and operational status data. Operational status data may refer to the dynamic state information of the flying vehicle during operation, including but not limited to flight mode, altitude, speed, acceleration, and heading, and is used to monitor the operating status of the flying vehicle in real time. Operational status data indicates the operational status of the flying vehicle, such as ground status or flight status. The designated airspace database may record a set of designated airspaces. The designated airspace database may store location information for each designated airspace within the set of designated airspaces. A designated airspace may be a pre-defined controlled area.

[0072] In practice, the designated airspaces stored in the designated airspace database may include specific airports and surrounding areas within a certain range, areas along specific boundaries, and areas containing important public infrastructure (e.g., power plants, substations, gas stations, water plants, public transportation hubs, avionics hubs, major water conservancy facilities, ports, highways, electrified railway lines, etc.). Optionally, the flying car's control system continuously outputs operating status data (e.g., a flight mode activation signal) to determine whether the flying car is in flight. If the flying car is in flight, the operating position data can be compared with the virtual coordinate information in the designated airspace database to determine whether the flying car is currently located within a preset sensitive area. Optionally, the designated airspace recorded in the designated airspace database can store virtual coordinate information corresponding to virtual coordinates that have undergone coordinate deflection. The operating position information can first be subjected to coordinate deflection processing to obtain the target virtual coordinates, and the target virtual coordinates are compared with the position coordinates of the designated airspace recorded in the designated airspace database. When it is identified that the flying car is in the designated airspace, the target operating data is deleted, wherein the target operating data is the operating data of the flying car in the designated airspace.

[0073] Optionally, when the flying car is in a land-based state, a pre-trained sensitive recognition model can be used for sensitive recognition processing. When the flying car is in flight, in addition to using the pre-trained sensitive recognition model for sensitive recognition processing, a designated airspace database will also be used for sensitive recognition processing to perform further desensitization processing.

[0074] Through this embodiment, by real-time monitoring of the flight status and operating position, it is ensured that the corresponding operating data is completely retained only when the flying car is in a non-sensitive area, effectively avoiding the improper collection and storage of sensitive information and enhancing the overall security of the data.

[0075] In an exemplary embodiment, the above method also includes: upon receiving an encrypted target data packet sent by a flying car, decrypting the encrypted target data packet through the cloud to obtain target data, wherein the target data includes desensitized data and other data in the operating data except sensitive data; obtaining the risk level corresponding to the specified data in the target data, wherein the specified data includes at least one of the following: point cloud data, ground information data; according to the risk level corresponding to the specified data, performing a third desensitization process on the specified data to obtain third desensitized data, and storing the third desensitized data in a preset database.

[0076] It should be noted that the cloud communicates with the flying car, and after the flying car generates a target data packet, the flying car will send an encrypted target data packet to the cloud with which it is communicating. The encrypted target data packet can be a data packet sent to the cloud through a preset secure transmission channel after the flying car performs initial desensitization and encryption operations on the vehicle side. The encrypted target data packet includes target data, which includes desensitized data and other data in the operating data except sensitive data. Among them, the desensitized data can be data that has been decrypted in the cloud and then desensitized on the vehicle side, such as personal sensitive information (face, license plate, etc.) has been replaced or blurred, and sensitive area data (coordinates, images, etc.) has also been deflected or deleted.

[0077] In addition to sensitive data, other data in the operational data can refer to data generated during the operation of the flying car that is not determined to be sensitive information, such as environmental perception data such as flight altitude, speed, acceleration, temperature, humidity, air pressure, and the vehicle's own status data (such as power, fuel level, system status, etc.). These data are crucial for the operation monitoring and data analysis of the flying car.

[0078] The designated data includes at least one of the following: point cloud data and geospatial data. Point cloud data can be three-dimensional environmental information collected by the flying car's lidar or other sensors, distributed in a cloud-like pattern of points. This data is used to construct a three-dimensional model of the flying car's surroundings, supporting functions such as obstacle detection, terrain understanding, and positioning and navigation. Geospatial data can refer to a collection of geographic environmental information and geophysical features recorded during the flying car's operation, also known as geographic information data. Optionally, geospatial data includes road attribute data and elevation data. Road attribute data can include geometric properties of the road (such as curvature, slope, and cross slope), as well as location information of facilities on and above the road. Elevation data can refer to the vertical height of the ground or a specific facility, i.e., the vertical distance relative to a reference plane (such as sea level). Geospatial data can be used for the flying car's path planning and environmental perception, and is critical data in the intelligent driving system.

[0079] Optionally, the process of obtaining the risk level corresponding to the specified data in the target data may be: performing risk level assessment processing on the point cloud data or the geospatial data through a preset model or a preset table to obtain the risk level corresponding to the specified data.

[0080] Optionally, a third desensitization process is performed on the designated data according to the risk level corresponding to the designated data to obtain third desensitized data. Specifically, the degree of desensitization of the designated data can be determined by the risk level corresponding to the designated data.

[0081] Optionally, in addition to the aforementioned desensitization processing module, the cloud can be equipped with a data preprocessing module to cleanse, format, handle missing values, detect and correct outliers, and standardize the target data received from the flying car to ensure data quality. Of course, the cloud can also be equipped with a data desensitization assessment module to evaluate the quality of the data desensitization through random checks, so that data that has been accidentally deleted or missed can be used to iterate the models deployed on the vehicle and in the cloud, continuously optimizing the desensitization effect.

[0082] Optionally, in practice, to further ensure the security of data processing in the cloud, a data center can be set up in the cloud. This data center includes a data compliance center (i.e., a compliance desensitization area) and a data business area. Internal and external networks are physically isolated, connected only to the compliance room via fiber optic cables. External access is through a VDI server to ensure data security. This ensures the secure flow of data between compliance processing and business use, preventing information leakage.

[0083] Similarly, to optimize data processing capabilities on the vehicle side, a basic management module can be configured on the vehicle side to provide SDK updates and version management services. To ensure secure communication between the vehicle and the cloud, preventing unauthorized access and data tampering, a vehicle authentication management module can be set up to handle vehicle authentication, certificate initialization, and updates.

[0084] This embodiment uses dual desensitization processing on both the vehicle and the cloud to completely eliminate or transform sensitive information from the flying car's operational data, reducing the risk of data leakage or misuse. The third desensitization process in the cloud is performed based on the data's risk level. This means that differentiated desensitization strategies can be adopted for data of different levels of sensitivity, protecting data security while avoiding excessive processing.

[0085] In an exemplary embodiment, a pre-trained risk level identification model is provided in the cloud; obtaining the risk level corresponding to specified data in the target data includes: inputting the specified data into the pre-trained risk level identification model to obtain the risk level corresponding to the specified data.

[0086] It should be noted that the pre-trained risk level identification model can be a machine learning model pre-trained in the cloud, specifically designed to identify and assess the risk level of specified data (such as point cloud data and geospatial data) within the flying car's operational data. The risk level identification model uses deep learning technology to learn the associations between various types of data and risk levels in historical data, and automatically performs risk assessments on newly input data.

[0087] In one example, the cloud receives an encrypted target data packet from a flying car. After decryption, the data preprocessing module first unifies the data format and performs quality checks to ensure that the data is suitable for further processing. Next, the target data is classified into five categories, for example, basic data, perception data, decision data, operating status data, and personal data. For specific data such as point cloud data and ground information data, the data processing engine uses this data as input and feeds it into a pre-trained risk level recognition model to identify the sensitivity of the point cloud data and ground information data. The specified data is then assigned a risk level, with the higher or lower level reflecting the sensitivity of the data. The risk level classification can include five levels, for example, DL1, DL2, DL3, DL4, and DL5.

[0088] Optionally, the training process for the risk level identification model may include: first, collecting a large amount of point cloud data and geospatial data from known sensitive area cases and non-sensitive area data as the basis for model training. Through feature engineering, feature data such as geographic attributes, density, and coverage of these point cloud data and geospatial data are extracted for model learning. Based on a deep learning framework, a suitable risk level identification algorithm, such as a convolutional neural network (CNN) or a long short-term memory network (LSTM), is selected to train the feature data. During training, cross-validation techniques are used to evaluate the model's generalization ability to ensure that it can accurately assess risk levels even on unseen data. Finally, through continuous iteration and optimization, including adjusting hyperparameters, increasing the amount of training data, and introducing data augmentation techniques, the accuracy and robustness of the model are continuously improved to ensure that it can effectively identify risk levels in flying car operation data.

[0089] Optionally, the cloud may also be provided with a preset level mapping table, which includes multiple preset types and preset risk levels corresponding to each of the multiple preset types; obtaining the risk level corresponding to specified data in the target data includes: according to the preset level mapping table, classifying and processing different types of operating data in the target data to obtain risk levels corresponding to different types of data in the target data, wherein the risk levels corresponding to different types of data in the target data include the risk level corresponding to the specified data.

[0090] Through this embodiment, through the pre-trained model, the cloud can automatically and quickly identify the risk level of designated data, avoiding the subjectivity and inefficiency of manual evaluation, and realizing the automation and intelligence of data risk assessment.

[0091] In an exemplary embodiment, a preset frame extraction table is provided in the cloud, and the preset frame extraction table includes a plurality of preset risk levels and a preset frame extraction strategy corresponding to one of the plurality of preset risk levels;

[0092] According to the risk level corresponding to the designated data, a third desensitizing process is performed on the designated data to obtain third desensitized data, including: when the designated data includes point cloud data, according to the risk level corresponding to the point cloud data, the frame extraction strategy corresponding to the point cloud data is determined from a preset frame extraction table; according to the frame extraction strategy corresponding to the point cloud data, the point cloud data is subjected to frame extraction process to obtain the point cloud data after frame extraction, and the third desensitized data includes the point cloud data after frame extraction.

[0093] It should be noted that the cloud can be configured with a preset frame extraction table. The preset frame extraction table may include a plurality of preset risk levels and a preset frame extraction strategy corresponding to one of the multiple preset risk levels. The preset frame extraction table can be used for point cloud data of different risk levels, and provides a preset frame extraction strategy. The frame extraction strategy indicates how to reduce the density of point cloud data in order to reduce the sensitivity of the data while maintaining the business value of the data as much as possible. Specifically, the frame extraction strategy may include the frame extraction ratio, the frame extraction algorithm, and the data format after frame extraction. The goal is to reduce the amount of sensitive information in the point cloud data without significantly affecting the data business performance.

[0094] Specifically, for high-risk point cloud data, a lower frame extraction frequency may be adopted to reduce the retention of detailed information; while for low-risk data, a higher frame extraction frequency may be maintained to retain more details. This risk-based frame extraction strategy solves the space and time cost issues during the storage and transmission of point cloud data. By properly adjusting the frame extraction frequency, it can protect sensitive information while maintaining data validity and integrity.

[0095] In one example, the cloud decrypts and identifies point cloud data within the target data. A risk level identification model evaluates the point cloud data and assigns a risk level based on the geographic area covered, point cloud density, and detailed information contained. Based on the risk level of the point cloud data, the cloud searches a preset abstraction table for the corresponding abstraction strategy. Assuming the data is assessed as high risk, the abstraction strategy may require retaining one point for every 100 points, significantly reducing data density and the resolution of sensitive information. The data processing engine processes the point cloud data according to the abstraction strategy, generating abstracted point cloud data.

[0096] This embodiment uses a pre-set frame extraction table to flexibly adjust the cloud's frame extraction strategy based on the different risk levels of the data. This ensures data security while avoiding over-processing of low-risk data, improving data utilization efficiency. The density of the point cloud data after frame extraction is reduced, reducing storage space requirements and data transmission bandwidth, and optimizing cloud resource utilization.

[0097] In an exemplary embodiment, the geospatial data includes road attribute data and elevation data, and a first preset table and a second preset table are provided on the cloud, wherein the first preset table includes a preset division gear corresponding to each of a plurality of preset risk levels; and the second preset table includes a preset fuzzy processing coefficient corresponding to each of a plurality of preset risk levels.

[0098] According to the risk level corresponding to the designated data, a third desensitization processing is performed on the designated data to obtain third desensitized data, including: when the designated data includes geospatial data, according to the risk level corresponding to the geospatial data, the division gear corresponding to the risk level is determined from the first preset table, and the fuzzy processing coefficient corresponding to the risk level is determined from the second preset table; according to the division gear, the road attribute data is graded to obtain the gear representation parameters corresponding to the road attribute data; according to the fuzzy processing coefficient, the elevation data is fuzzified to obtain the fuzzified data corresponding to the elevation data, wherein the third desensitized data includes the gear representation parameters corresponding to the road attribute data and the fuzzified data corresponding to the elevation data.

[0099] It should be noted that the designated data may include geospatial data, which may include road attribute data and elevation data. Road attribute data may refer to a series of characteristic data related to the road, such as the road's material, width, slope, curvature, cross slope, road signs, etc.

[0100] Elevation data can refer to the height of the ground or a specific reference point. For flying cars, elevation data helps to accurately understand and avoid obstacles and ensure safe flight.

[0101] The cloud deploys a first preset table and a second preset table. The first preset table defines the road attribute data levels for each preset risk level. Level representation parameters convert road attribute data into discrete values to reduce the sensitivity and confidentiality of the data.

[0102] The second preset table can be used to provide a fuzzy processing coefficient for each preset risk level to guide the fuzzy processing of elevation data to reduce the exposure of precise location information while maintaining basic usability of the data.

[0103] In one example, the risk level identification model assesses the risk level of geospatial data. For example, geospatial data covering sensitive areas or containing critical infrastructure information may be labeled high risk. Based on this risk level, the cloud searches the corresponding classification level from a first preset table and converts the road attribute data into several preset level representation parameters. For example, slope may be classified into three levels: "low," "medium," and "high," while road width may be classified into "narrow," "medium," and "wide," reducing the risk of leaking precise values. Simultaneously, a fuzzification coefficient is obtained from a second preset table to fuzzify the elevation data. The coefficient determines the degree of fuzzification, with higher risk levels corresponding to higher fuzziness to ensure the security of sensitive location information. For example, for a high risk level, elevation data may be fuzzified to a tolerance of plus or minus 5 meters, making the specific location information unclear while still supporting the flying car's environmental understanding.

[0104] Specifically, geospatial data processing is achieved through the use of scales and fuzzification coefficients. The scales simplify road attribute data, such as dividing road widths into several standard intervals, while the fuzzification coefficients reduce the precision of elevation data, such as by adding noise or smoothing to blur actual height values. This processing method allows precise control of data sensitivity by adjusting scales and fuzzification coefficients, while also preserving the essential geographic characteristics of the geospatial data to a certain extent.

[0105] Through this embodiment, by means of the grading processing of the road attribute data and the fuzzification processing of the elevation data, the road detail information involving sensitive areas is effectively protected and the leakage of information is avoided.

[0106] It should be noted that the acquisition, storage, use, and processing of data in the above embodiments of this application are in compliance with the relevant provisions of national laws and regulations.

[0107] Figure 4 is a schematic diagram of another optional data processing method in this optional example, such as Figure 4 As shown, the data processing process may include:

[0108] Step 1: On the flying car side, basic management and vehicle authentication management are required. This basic management includes SDK updates and SDK version management. SDK updates ensure the latest version of the software development kit (SDK), providing the latest data processing and compliance control features. SDK version management is used to track and manage different versions of the software package, ensuring consistency and promptly resolving known issues. Vehicle authentication management provides vehicle identity authentication services, ensuring the legality of the flying car and ensuring data compliance. It also manages and maintains the digital certificates required for communication between the flying car and the cloud, ensuring data transmission security.

[0109] Step 2: When the basic management and vehicle authentication management are passed, coordinate deflection processing is performed on the operating position data in the operating data to obtain virtual coordinates.

[0110] In step 3, once the virtual coordinates corresponding to the operational location data are obtained, the pre-trained sensitive area recognition model is used to identify the operational data and desensitize the identified sensitive area data, regardless of whether the flying car is in a land-based or flight state. When the flying car is in flight, the designated airspace database is used to monitor whether the vehicle is in a designated airspace and automatically delete the operational data involving the designated airspace.

[0111] Step 4: When the flying car is in a land or flight state, the running image is processed by a pre-trained image recognition model. If the image recognition result indicates the presence of personal sensitive data, the personal sensitive data is desensitized, for example, by blocking, blurring, or replacing the data.

[0112] Step 5: Use a preset algorithm to encrypt the target data to obtain an encrypted target data packet, where the target data includes the desensitized data and other data in the running data except the sensitive data.

[0113] Step 6: During data transmission, the encryption key is automatically distributed and regularly updated through the key management system to ensure the security of data decryption and subsequent processing in the cloud.

[0114] Step 7: Perform two-way authentication through the key management system to enhance the security of the data transmission channel between the cloud and the flying car.

[0115] Step 8: Distribute the target data packet to the cloud.

[0116] Step 9: decrypt the encrypted target data packet.

[0117] Step 10: perform data preprocessing.

[0118] Among them, data preprocessing includes cleaning, format unification, missing value processing, outlier detection and correction, data standardization and other operations to ensure data quality and consistency.

[0119] Step 11: classify and grade the target data.

[0120] Specifically, target data is categorized into five categories: basic data, perception data, decision-making data, operational status data, and personal data. For specific data, such as point cloud data and geospatial data, the data processing engine uses this data as input and feeds it into a pre-trained risk level recognition model to identify the sensitivity of the point cloud and geospatial data. This model then assigns a risk level, reflecting the data's sensitivity. Risk levels can be divided into five levels: DL1, DL2, DL3, DL4, and DL5.

[0121] In step 12, the classified target data can be initially stored.

[0122] Step 13: extract the frame of the point cloud data in the target data after classification and grading to reduce the data density and sensitive information.

[0123] Step 14: desensitize the geospatial data.

[0124] Specifically, the geospatial data includes road attribute data and elevation data, and a first preset table and a second preset table are set in the cloud, wherein the first preset table includes a preset division gear corresponding to each preset risk level in multiple preset risk levels; the second preset table includes a preset fuzzy processing coefficient corresponding to each preset risk level in multiple preset risk levels; according to the risk level corresponding to the geospatial data, the division gear corresponding to the risk level is determined from the first preset table, and the fuzzy processing coefficient corresponding to the risk level is determined from the second preset table; according to the division gear, the road attribute data is graded to obtain the gear representation parameters corresponding to the road attribute data; according to the fuzzy processing coefficient, the elevation data is fuzzified to obtain fuzzy data corresponding to the elevation data.

[0125] Step 15: Data desensitization check;

[0126] Specifically, spot checks are conducted to determine the effectiveness of data desensitization, assess the quality of compliance processing, and ensure data compliance. The cloud can be equipped with a data desensitization assessment module to evaluate the quality of data desensitization through random checks. This allows for the use of accidentally deleted or missed data in model iterations deployed on the vehicle and in the cloud, continuously optimizing the desensitization effect.

[0127] Step 16: storing the desensitized target data, and using data compression during the storage process;

[0128] Step 17: Data distribution.

[0129] Specifically, according to business needs, the stored data is decompressed and distributed to the corresponding application system to support intelligent driving and other data-driven functions. Through this optional example, a comprehensive data security protection system is built through the collaboration of vehicle-side pre-processing and cloud-side deep processing. The instant identification, desensitization and encryption technology of sensitive data on the vehicle side effectively reduce the risk of exposure of sensitive information during transmission. At the same time, vehicle authentication ensures the legitimacy of the data source. Intelligent classification, point cloud extraction and geospatial data fuzzification processing in the cloud protect geographically sensitive information. Through the combination of automation and manual spot checks, compliance effects are continuously optimized. Full-link encryption and two-way authentication enhance the security of data transmission, and the key management mechanism of KMS avoids loopholes caused by human factors. The application of data preprocessing and compression technology ensures data quality while optimizing storage and transmission efficiency.

[0130] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0131] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM (Read-Only Memory, Read-Only Memory) / RAM (Random Access Memory, Random Access Memory), a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0132] According to another aspect of the embodiments of the present application, a data processing device is further provided, which can be used to implement the data processing method provided in the above embodiments, and will not be repeated hereafter. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0133] Figure 5is a structural block diagram of an optional data processing device according to an embodiment of the present application, such as Figure 5 As shown in , the data processing device includes:

[0134] an acquisition unit 502 for acquiring operation data of the flying car while the flying car is in operation, wherein the flying car communicates with the cloud and a pre-trained sensitive recognition model is deployed on the flying car;

[0135] The identification unit 504 is used to perform sensitive identification processing on the operating data using a sensitive identification model to obtain a sensitive identification result;

[0136] A first desensitizing unit 506 is configured to desensitize the sensitive data to obtain desensitized data when the sensitivity identification result indicates that the operation data contains sensitive data;

[0137] The encryption unit 508 is used to encrypt the desensitized data and other data in the running data except the sensitive data, and send the encrypted target data packet to the cloud.

[0138] It should be noted that the acquisition unit 502 in this embodiment can be used to execute the above step S202, the identification unit 504 in this embodiment can be used to execute the above step S204, the first desensitizing unit 506 in this embodiment can be used to execute the above step S206, and the encryption unit 508 in this embodiment can be used to execute the above step S208.

[0139] Through the embodiments provided by the present application, when a flying car is running, a pre-trained sensitive identification model deployed on the flying car is used to perform sensitive identification on the operating data, and when the sensitive identification result indicates the presence of sensitive data, the sensitive data is desensitized, and then the desensitized data and other data in the operating data except the sensitive data are encrypted to transmit the encrypted target data packet to the cloud, thereby reducing the risk of sensitive information leakage before the data is transmitted to the cloud, and to a certain extent solving the problem of relatively high leakage risk during transmission of original operating data in related technologies, improving the security of data processing, and also reducing the data processing burden on the cloud.

[0140] In one exemplary embodiment, there are multiple sensitive recognition models, including a sensitive area recognition model and an image recognition model. The flying vehicle is equipped with at least one visual perception device, and the operational data includes at least one visual perception device used to capture operational images of the flying vehicle during operation. Recognition unit 504 is further configured to: input the operational data into the sensitive area recognition model and output a predicted recognition result, wherein the predicted recognition result indicates whether sensitive area data exists in the operational data; input the operational image into the image recognition model and output an image recognition result corresponding to the operational image, wherein the image recognition result indicates whether personal sensitive data exists in the operational image. Personal sensitive data includes at least one of the following: facial information and license plate information.

[0141] In an exemplary embodiment, the first desensitizing unit 506 is also used to, when the sensitive data is sensitive area data, perform a first desensitizing process on the sensitive area data to obtain first desensitized data, wherein the first desensitizing process includes at least one of the following: deletion process, replacement process; and when the sensitive data is personal sensitive data, perform a second desensitizing process on the personal sensitive data to obtain second desensitized data, wherein the second desensitizing process includes at least one of the following: blocking process, blurring process, replacement process.

[0142] In an exemplary embodiment, the operating data includes operating position data; the above-mentioned device also includes: a deflection unit, which is used to perform coordinate deflection processing on the operating position data to obtain virtual coordinate data corresponding to the operating position data, wherein the coordinate deflection processing includes at least one of the following: coordinate conversion processing, coordinate mapping processing.

[0143] In an exemplary embodiment, the operating data includes operating position data and operating status data. The flying car is deployed with a designated airspace database, which can record a set of designated airspaces. The above-mentioned device also includes: a deletion unit, which is used to: when the operating status data indicates that the flying car is in a flying state, identify whether the flying car is in a designated airspace in the designated airspace database based on the operating position data; when the flying car is in the designated airspace, delete the target operating data, wherein the target operating data is the operating data of the flying car in the designated airspace.

[0144] In an exemplary embodiment, the above-mentioned device also includes: a second desensitizing unit, which is used to decrypt the encrypted target data packet sent by the flying car through the cloud to obtain target data, wherein the target data includes desensitized data and other data in the operating data except sensitive data; obtain the risk level corresponding to the specified data in the target data, wherein the specified data includes at least one of the following: point cloud data, ground information data; according to the risk level corresponding to the specified data, perform a third desensitizing process on the specified data to obtain third desensitized data, and store the third desensitized data in a preset database.

[0145] In an exemplary embodiment, a pre-trained risk level identification model is provided in the cloud; the second desensitizing unit is further used to: input the specified data into the pre-trained risk level identification model to obtain the risk level corresponding to the specified data.

[0146] In an exemplary embodiment, a preset frame extraction table is provided in the cloud, and the preset frame extraction table includes a plurality of preset risk levels and a preset frame extraction strategy corresponding to one of the plurality of preset risk levels.

[0147] The second desensitizing unit is further used for: when the designated data includes point cloud data, determining the frame extraction strategy corresponding to the point cloud data from a preset frame extraction table according to the risk level corresponding to the point cloud data; performing frame extraction processing on the point cloud data according to the frame extraction strategy corresponding to the point cloud data to obtain the point cloud data after frame extraction. The third desensitized data includes the point cloud data after frame extraction.

[0148] In an exemplary embodiment, the geospatial data includes road attribute data and elevation data, and a first preset table and a second preset table are provided in the cloud, wherein the first preset table includes a preset division gear corresponding to each preset risk level in a plurality of preset risk levels; and the second preset table includes a preset fuzzy processing coefficient corresponding to each preset risk level in a plurality of preset risk levels.

[0149] The second desensitizing unit is further used for: when the designated data includes geospatial data, determining the division gear corresponding to the risk level from the first preset table according to the risk level corresponding to the geospatial data, and determining the fuzzy processing coefficient corresponding to the risk level from the second preset table; performing division processing on the road attribute data according to the division gear to obtain the gear representation parameters corresponding to the road attribute data; performing fuzzy processing on the elevation data according to the fuzzy processing coefficient to obtain the fuzzified data corresponding to the elevation data, wherein the third desensitized data includes the gear representation parameters corresponding to the road attribute data and the fuzzified data corresponding to the elevation data.

[0150] It should be noted that the above modules can be implemented through software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0151] It should be noted that the acquisition, storage, use, and processing of data in the above embodiments of this application are in compliance with the relevant provisions of national laws and regulations.

[0152] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored program, wherein the program executes the steps of any of the above method embodiments when it is run.

[0153] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, a ROM, a RAM, a mobile hard disk, a magnetic disk, or an optical disk.

[0154] According to another aspect of the embodiments of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to execute the steps of any of the above-described method embodiments through the computer program. In an exemplary embodiment, the electronic device may further comprise a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0155] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail here.

[0156] According to another aspect of an embodiment of the present application, a computer program product is also provided, which includes a computer program / instruction, which contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication portion 609, and / or installed from the removable medium 611. When the computer program is executed by the central processing unit 601, the various functions provided by the embodiments of the present application are performed. The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0157] Figure 6 The following schematically shows a block diagram of a computer system structure of an electronic device for implementing an embodiment of the present application. Figure 6As shown, computer system 600 includes a CPU (Central Processing Unit) 601, which can perform various appropriate actions and processes according to programs stored in ROM 602 or programs loaded from storage unit 608 into RAM 603. Various programs and data required for system operation are also stored in random access memory 603. CPU 601, read-only memory 602, and random access memory 603 are connected to each other via bus 604. I / O (Input / Output) interface 605 is also connected to bus 604.

[0158] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, and the like; an output section 607 including devices such as a CRT (Cathode Ray Tube), an LCD (Liquid Crystal Display), and speakers; a storage section 608 including a hard disk; and a communication section 609 including a network interface card such as a local area network card or a modem. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output interface 605 as needed. Removable media 611, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 610 as needed, so that computer programs read therefrom can be installed into the storage section 608 as needed.

[0159] In particular, according to an embodiment of the present application, the processes described in the various method flow charts can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for executing the methods shown in the flow charts. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609 and / or installed from a removable medium 611. When the computer program is executed by the central processing unit 601, the various functions defined in the system of the present application are performed.

[0160] It should be noted that Figure 6 The computer system 600 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0161] Obviously, those skilled in the art should understand that the modules or steps of the present application described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices, they can be implemented using program code executable by the computing device, and thus, they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be performed in a different order than herein, or they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0162] The above are only preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A data processing method, characterized in that: include: Acquiring operational data of a flying car while the flying car is in operation, wherein the flying car communicates with a cloud and a pre-trained sensitive recognition model is deployed on the flying car; Performing sensitive identification processing on the operating data using the sensitive identification model to obtain a sensitive identification result; If the sensitivity identification result indicates that sensitive data exists in the operation data, desensitizing the sensitive data to obtain desensitized data; The desensitized data and other data in the operating data except the sensitive data are encrypted, and the encrypted target data packet is sent to the cloud.

2. The method according to claim 1, characterized in that There are multiple sensitive recognition models, the multiple sensitive recognition models include sensitive area recognition models and image recognition models, the flying car is provided with at least one visual perception device, and the operation data includes the at least one visual perception device used to collect operation images of the flying car during operation; The performing sensitive identification processing on the operating data by using the sensitive identification model to obtain a sensitive identification result includes: Inputting the operating data into the sensitive area recognition model and outputting a prediction recognition result, wherein the prediction recognition result is used to indicate whether sensitive area data exists in the operating data; The running image is input into the image recognition model, and an image recognition result corresponding to the running image is output, wherein the image recognition result is used to indicate whether personal sensitive data exists in the running image, and the personal sensitive data includes at least one of the following: facial information, license plate information.

3. The method according to claim 2, characterized in that The desensitizing the sensitive data to obtain desensitized data includes: In the case where the sensitive data is the sensitive area data, performing a first desensitization process on the sensitive area data to obtain first desensitized data, wherein the first desensitization process includes at least one of the following: deletion process, replacement process; In the case where the sensitive data is the personal sensitive data, the personal sensitive data is subjected to a second desensitizing process to obtain second desensitized data, wherein the second desensitizing process includes at least one of the following: blocking processing, blurring processing, and replacement processing.

4. The method according to claim 1, wherein The operation data includes operation position data; the method further includes: The operating position data is subjected to coordinate deflection processing to obtain virtual coordinate data corresponding to the operating position data, wherein the coordinate deflection processing includes at least one of the following: coordinate conversion processing and coordinate mapping processing.

5. The method according to claim 1, characterized in that The operation data includes operation position data and operation status data. The flying car is deployed with a designated airspace database, and the designated airspace database records a set of designated airspaces. The method further includes: When the operating state data indicates that the flying car is in a flying state, identifying whether the flying car is in a designated airspace in a designated airspace database according to the operating position data; When the flying car is in the designated airspace, target operation data is deleted, wherein the target operation data is the operation data of the flying car in the designated airspace.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: Upon receiving the encrypted target data packet sent by the flying car, decrypting the encrypted target data packet through the cloud to obtain target data, wherein the target data includes the desensitized data and other data in the operating data except the sensitive data; Obtaining a risk level corresponding to designated data in the target data, wherein the designated data includes at least one of the following: point cloud data, geospatial data; According to the risk level corresponding to the designated data, a third desensitization process is performed on the designated data to obtain third desensitized data, and the third desensitized data is stored in a preset database.

7. The method according to claim 6, characterized in that The cloud is provided with a pre-trained risk level identification model; The obtaining of the risk level corresponding to the specified data in the target data includes: The designated data is input into the pre-trained risk level identification model to obtain the risk level corresponding to the designated data.

8. The method according to claim 6, characterized in that The cloud is provided with a preset frame extraction table, wherein the preset frame extraction table includes a plurality of preset risk levels and a preset frame extraction strategy corresponding to one of the plurality of preset risk levels; The performing a third desensitization process on the designated data according to the risk level corresponding to the designated data to obtain third desensitized data includes: In a case where the designated data includes point cloud data, determining a frame extraction strategy corresponding to the point cloud data from the preset frame extraction table according to a risk level corresponding to the point cloud data; According to the frame extraction strategy corresponding to the point cloud data, the point cloud data is subjected to frame extraction processing to obtain the point cloud data after frame extraction, and the third desensitized data includes the point cloud data after frame extraction.

9. The method according to claim 6, characterized in that The geospatial data includes road attribute data and elevation data, and the cloud is provided with a first preset table and a second preset table, wherein the first preset table includes a preset division gear corresponding to each preset risk level in a plurality of preset risk levels; and the second preset table includes a preset fuzzy processing coefficient corresponding to each preset risk level in a plurality of preset risk levels; The performing a third desensitization process on the designated data according to the risk level corresponding to the designated data to obtain third desensitized data includes: In a case where the designated data includes the geospatial data, determining, according to the risk level corresponding to the geospatial data, a division level corresponding to the risk level from the first preset table, and determining a fuzzy processing coefficient corresponding to the risk level from the second preset table; According to the gear division, the road attribute data is subjected to gear division processing to obtain gear representation parameters corresponding to the road attribute data; According to the fuzzy processing coefficient, the elevation data is fuzzified to obtain fuzzy data corresponding to the elevation data, wherein the third desensitized data includes the gear representation parameters corresponding to the road attribute data and the fuzzy data corresponding to the elevation data.

10. A data processing device, characterized in that: include: an acquisition unit, configured to acquire operating data of the flying car while the flying car is in operation, wherein the flying car communicates with the cloud and a pre-trained sensitive recognition model is deployed on the flying car; an identification unit, configured to perform sensitive identification processing on the operating data using the sensitive identification model to obtain a sensitive identification result; a first desensitizing unit, configured to, when the sensitivity identification result indicates that sensitive data exists in the operation data, perform desensitization processing on the sensitive data to obtain desensitized data; An encryption unit is used to encrypt the desensitized data and other data in the operating data except the sensitive data, and send the encrypted target data packet to the cloud.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the method according to any one of claims 1 to 9 when executed by a processor.

12. 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 computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images