Data processing method, storage medium, electronic device and program product

Abstract

The invention provides a data processing method, a storage medium, electronic equipment and a program product, and relates to the technical field of data as a service. The data processing method comprises the steps of determining a trial acquisition parameter set based on a data acquisition demand of a user; on the basis of the pilot production parameter set, pilot production operation is executed, and pilot production data of the intelligent equipment with the body are obtained; if the trial production data passes the quality verification, generating a data acquisition specification based on the trial production parameter set; and executing mass production acquisition operation based on the data acquisition specification to obtain target acquisition data of the intelligent device with the body. According to the method, the data acquisition specification is dynamically generated according to the quality verification result of the trial acquisition data, so that the accuracy and effectiveness of the obtained target acquisition data are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data-as-a-service, and particularly relates to a data processing method, a storage medium, an electronic device and a program product. BACKGROUND

[0002] As a new data management and delivery mode, data-as-a-service (DaaS) is gradually changing the traditional data utilization mode in various industries. For example, for embodied intelligence industry, the use scenario of embodied intelligence equipment is usually preset, and the embodied intelligence equipment is used for data collection based on the required collection parameters in the use scenario to obtain the collection data of the embodied intelligence equipment.

[0003] However, since the collection parameters corresponding to the preset use scenario are fixed, the diversified collection requirements of users cannot be met, thereby affecting the accuracy and effectiveness of the collection data of the embodied intelligence equipment. SUMMARY

[0004] Therefore, the embodiments of the present application provide a data processing method, a storage medium, an electronic device and a program product.

[0005] In a first aspect, an embodiment of the present application provides a data processing method, comprising: determining a trial collection parameter set based on a data collection requirement of a user; performing a trial collection operation based on the trial collection parameter set to obtain trial collection data of embodied intelligence equipment; if the trial collection data passes quality verification, generating a data collection specification based on the trial collection parameter set; and performing mass production collection operation based on the data collection specification to obtain target collection data of the embodied intelligence equipment.

[0006] In combination with the first aspect, in some implementation manners of the first aspect, generating the data collection specification based on the trial collection parameter set comprises: generating a feedback coefficient based on the trial collection parameter set; calculating an environment quantization index based on an environmental dynamic parameter and a collection task complexity parameter; fusing the environment quantization index and a collection task difficulty index based on the feedback coefficient to obtain a sensor configuration intensity value, the sensor configuration intensity value being used to represent at least one index in a type of required sensor, a sampling frequency and a spatial resolution; and generating the data collection specification based on the sensor configuration intensity value.

[0007] In combination with the first aspect, in some implementation manners of the first aspect, the method further comprises: when performing the mass production collection operation, at least one of the following monitoring is further performed: monitoring whether the point cloud density collected by the laser radar of the embodied intelligence equipment is greater than a target density; monitoring whether the data collection interval of the inertial measurement unit of the embodied intelligence equipment is less than a target time; and monitoring whether an action sequence performed by the embodied intelligence equipment is complete.

[0008] In combination with the first aspect, in certain implementations of the first aspect, after performing mass production acquisition operations based on data acquisition specifications to obtain target acquisition data of the embodied smart device, it also includes: performing a packaging operation on the target acquisition data; wherein the packaging operation includes at least one of the following items: converting the target acquisition data into a standard format to reduce the adaptation cost between processing algorithms corresponding to the target acquisition data; compressing the target acquisition data to reduce the amount of data; generating metadata corresponding to the target acquisition data, the metadata including scene parameters and / or sensor calibration parameters.

[0009] In combination with the first aspect, in certain implementations of the first aspect, after performing mass production collection operations based on data collection specifications to obtain target collection data of the embodied smart device, it also includes: performing privacy protection operations on the target collection data; wherein the privacy protection operations include at least one of the following items: desensitizing the target collection data; generating key information bound to the target collection data to encrypt and store and / or decrypt and access the target collection data; using blockchain to record access rights to the target collection data to verify the access rights to the target collection data.

[0010] In combination with the first aspect, in certain implementations of the first aspect, after performing mass production collection operations based on data collection specifications to obtain target collection data of the embodied smart device, it also includes: based on the data delivery method selected by the user, selecting at least one from application programming interface, file transfer and cloud storage to send the target collection data to the user.

[0011] In combination with the first aspect, in some implementations of the first aspect, it also includes: if the trial sampling data fails the quality verification, sending parameter adjustment suggestions to the user; in response to the user's confirmation operation of the parameter adjustment suggestion, generating a new trial sampling parameter set; based on the new trial sampling parameter set, re-acquiring the trial sampling data of the embodied smart device, and performing quality verification on the re-acquired trial sampling data.

[0012] In a second aspect, an embodiment of the present application provides a data processing device, including:

[0013] The parameter set determination module is used to determine the test sampling parameter set based on the user's data collection requirements;

[0014] A test mining operation module is used to perform a test mining operation based on a test mining parameter set and obtain test mining data of the embodied intelligent device;

[0015] A specification generation module is used to generate data collection specifications based on the test collection parameter set if the test collection data passes the quality verification;

[0016] The mass production acquisition module is used to perform mass production acquisition operations based on data acquisition specifications to obtain target acquisition data of embodied smart devices.

[0017] In a third aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program for executing the method in the first aspect.

[0018] In a fourth aspect, an embodiment of the present application provides an electronic device, which comprises: a processor; a memory for storing processor-executable instructions; and the processor is configured to execute the method in the first aspect.

[0019] In a fifth aspect, an embodiment of the present application provides a computer program product, which comprises instructions for causing an electronic device to implement the method in the first aspect when the instructions are executed on the electronic device.

[0020] In the present application, based on the data collection requirements of a user, a trial collection parameter set is determined, a trial collection operation is performed based on the trial collection parameter set, and trial collection data of the embodied intelligent device is obtained, so that the trial collection parameters used for data collection can meet different collection requirements of the user. Moreover, if the trial collection data passes the quality verification, the data collection specification is generated based on the trial collection parameter set, so as to ensure the scientificity and reliability of the specification, avoid the specification deviation caused by the quality problem of the trial collection data, and reduce the data correction cost in the mass production stage. Finally, the mass production collection operation is performed based on the data collection specification, and the target collection data of the embodied intelligent device is obtained, which is conducive to improving the accuracy and effectiveness of the obtained target collection data. BRIEF DESCRIPTION OF DRAWINGS

[0021] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which:

[0022] Figure 1 Fig. 1 shows a flowchart of a data processing method according to an embodiment of the present application.

[0023] Figure 2 Fig. 2 shows a flowchart of a trial collection parameter set adjustment process according to an embodiment of the present application.

[0024] Figure 3 Fig. 3 shows an architecture diagram of a data processing system according to an embodiment of the present application.

[0025] Figure 4 Fig. 4 shows a flowchart of an operation standard operating procedure fusion scheme according to an embodiment of the present application.

[0026] Figure 5 Fig. 1 shows a structural schematic diagram of a data processing apparatus provided by an embodiment of the present application.

[0027] Figure 6 Fig. 2 shows a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0029] Firstly, the technical solution can be applied to various scenes requiring data collection of embodied intelligent devices, such as data collection of autonomous vehicles, data collection of robot working environments, etc. The collected data can be used for subsequent data analysis, model training or decision support, etc. In the application scene of autonomous vehicles, by collecting various sensor data in the driving process of the vehicle, real-time monitoring and analysis of the driving state of the vehicle can be realized, and then the autonomous driving algorithm can be optimized to improve the safety and reliability of autonomous driving. In the data collection scene of robot working environments, by collecting environmental data and sensor data in the working process of the robot, accurate grasp of the working state of the robot can be realized, and then the working tasks of the robot can be more reasonably planned and scheduled.

[0030] In these scenes, according to the specific environmental conditions of the embodied intelligent device, different parameters need to be collected. In the prior art, standardized data collection is performed by using the collection parameters corresponding to the use scenarios in the pre-defined scenario library, which leads to the inability to dynamically respond to the adjustment requirements of the collection parameters caused by the changes of the environmental conditions, the solidification of the collection parameters, and the lack of flexibility.

[0031] The technical solution can determine a trial collection parameter set based on the data collection requirements of the user, dynamically generate a data collection specification through the quality verification result of the trial collection data, and then improve the accuracy and effectiveness of the target collection data of the embodied intelligent device.

[0032] Figure 1 Fig. 3 shows a flowchart of a data processing method provided by an embodiment of the present application. The embodiment can be applied to the case where a data collection specification is determined, and a data collection operation is performed on an embodied intelligent device according to the data collection specification to obtain target collection data. As shown in the figure, Figure 1 The method comprises the following steps.

[0033] Step S110, determining a set of trial collection parameters based on the data collection requirement of the user.

[0034] The data collection requirement is a specific data collection requirement proposed by the user for the embodied intelligent device. The data collection requirement can be a requirement for a collection scene, for example, whether the scene to be collected is indoors or outdoors, daytime or nighttime. It can also be a requirement for the type of collected data, for example, whether image data, sound data or other types of data need to be collected. It can also be a requirement for the quality of collected data, for example, requirements for resolution, sampling frequency, etc.

[0035] The set of trial collection parameters is a set of parameters for trial collection operation determined based on the data collection requirement. For example, when the embodied intelligent device is a robot, the parameters in the set of trial collection parameters can be the type of sensor, sampling frequency, and spatial resolution, etc. It can also be the rotation angle of the joint of the mechanical arm, the position of the end effector, the gripping force of the mechanical arm, and the movement speed of the mechanical arm, etc.

[0036] In a specific implementation, in response to the information input operation of the scene requirement information, the data collection requirement is obtained by analyzing the input scene requirement information. Exemplarily, the input scene requirement information can be in the form of natural language, and the input scene requirement information is analyzed by natural language processing (NLP) technology. The NLP is an artificial intelligence technology that can understand and analyze human natural language.

[0037] In this embodiment, after obtaining the data collection requirement, the specific implementation of determining the set of trial collection parameters based on the data collection requirement can include the following: 1. Pre-set the mapping relationship between various data collection requirements and corresponding trial collection parameters. After obtaining the data collection requirement, the parameters matching the data collection requirement are determined based on the mapping relationship, and the set of trial collection parameters is composed of the determined parameters. 2. Input the data collection requirement into a pre-trained parameter determination model to output the corresponding set of trial collection parameters. The parameter determination model can be a machine learning model.

[0038] Step S120, performing a trial collection operation based on the set of trial collection parameters to obtain trial collection data of the embodied intelligent device.

[0039] In this embodiment, for each parameter in the set of trial collection parameters, the data corresponding to the parameter in the working process of the embodied intelligent device is collected to form the trial collection data. It should be noted that the trial collection operation is for the collection of a small amount of data. The trial collection process is mainly to verify the parameters in the set of trial collection parameters, and the trial collection data is not the final data to be delivered. Therefore, in order to improve the data processing efficiency, a small amount of data can be collected as trial collection data.

[0040] Exemplarily, the first preset amount of trial production data can be collected through the trial production operation, or the trial production data generated by the body intelligent device within the first preset time length can be collected through the trial production operation. The first preset amount and the first preset time length can be set by the person skilled in the art according to the actual application, and the present embodiment does not limit this.

[0041] In step S130, if the trial production data passes the quality verification, the data collection specification is generated based on the trial production parameter set.

[0042] In the present embodiment, the quality verification can be at least one of missing data verification, data anomaly verification and noise verification. The missing data verification is used to verify whether the trial production data is missing key information. If the key information is missing, the missing data verification fails; otherwise, the missing data verification passes. The data anomaly verification is used to verify whether the data value is within the preset value range. If yes, the data anomaly verification passes; otherwise, the data anomaly verification fails. The noise verification is used to verify whether the noise amount exceeds the preset noise amount. If yes, the noise verification fails; otherwise, the noise verification passes.

[0043] It should be noted that, in order to ensure the accuracy of the parameters used for collecting data, in the case where the quality verification includes missing data verification, data anomaly verification and noise verification, only when each verification passes, it means that the quality verification passes; if at least one verification fails, it means that the quality verification fails.

[0044] In the present embodiment, if the trial production data passes the quality verification, it means that the trial production parameter set can better meet the data collection needs of the user, and therefore, the data collection specification for subsequent mass production collection operation can be generated based on the trial production parameter set.

[0045] It can be understood that the main difference between the trial production parameter set and the data collection specification is that the trial production parameter set is a set of parameters set for the trial production operation, mainly used for preliminary verification of the parameters. The data collection specification is generated based on the trial production parameter set that passes the verification, and is used to guide the subsequent mass production collection operation specification.

[0046] Exemplarily, the data collection specification can be directly composed of the parameters in the trial collection parameter set and other data collection requirements, for example, the other data collection requirements include at least one of a time range of data collection, a data storage format, a data storage location, and a transmission manner adopted when delivering data. In addition, in order to collect more accurate and high-quality data in the mass production stage, the parameters in the trial collection parameter set can be further refined, and the refined trial collection parameter set and the other data collection requirements are used to generate the data collection specification. For example, for the robot, the sampling frequency of the camera in the trial collection parameter set is required to be any value between 30 frames per second and 60 frames per second. After refinement, the sampling frequency of the camera is required to be any value between 40 frames per second and 50 frames per second, and the data collection specification is composed of the refined parameters. Through refinement, the data collection specification is more in line with the actual application requirements, which is beneficial to improve the accuracy and effectiveness of the collected data.

[0047] In step S140, the mass production collection operation is performed based on the data collection specification, and target collection data of the embodied intelligent device is obtained.

[0048] In this embodiment, the embodied intelligent device can be subjected to the mass production collection operation according to the data collection specification. Exemplarily, the mass production collection operation can collect data of a second preset data amount; or collect data generated by the embodied intelligent device within a second preset time length. The second preset data amount is greater than the first preset data amount collected in the trial collection operation; and the second preset time length is greater than the first preset time length corresponding to the trial collection operation. The data obtained after the mass production collection operation is performed is used as the target collection data of the embodied intelligent device.

[0049] In this embodiment, the trial collection parameter set is determined based on the data collection requirements of the user, the trial collection operation is performed based on the trial collection parameter set, and trial collection data of the embodied intelligent device is obtained, so that the trial collection parameters used for data collection can meet different collection requirements of the user. In addition, if the trial collection data passes the quality verification, the data collection specification is generated based on the trial collection parameter set, so as to ensure the scientificity and reliability of the specification, avoid specification deviation caused by quality problems of the trial collection data, and reduce the data correction cost in the mass production stage. Finally, the mass production collection operation is performed based on the data collection specification, and the target collection data of the embodied intelligent device is obtained, which is beneficial to improve the accuracy and effectiveness of the obtained target collection data.

[0050] Optionally, the specific implementation manner of generating the data collection specification based on the trial collection parameter set can be: generating a feedback coefficient based on the trial collection parameter set; calculating an environment quantization index based on the environmental dynamic parameter and the collection task complexity parameter; fusing the environment quantization index and the collection task difficulty index based on the feedback coefficient to obtain a sensor configuration intensity value, the sensor configuration intensity value being used to represent at least one index in a type of required sensor, a sampling frequency, and a spatial resolution; and generating the data collection specification based on the sensor configuration intensity value.

[0051] Optionally, the feedback coefficient can be determined based on a quality evaluation result of the trial collection data. For example, the higher the quality of the trial collection data, the greater the feedback coefficient can be set to represent that the trial collection parameter set is more matched with the data collection demand of the user and the environmental condition.

[0052] Specifically, the feedback coefficient can be generated based on the trial collection data corresponding to the trial collection parameter set. For example, the dynamic data stream composed of the data collected by the sensor at different collection time points in the trial collection data is taken as a real-time feedback signal, and a sine function value is taken for the real-time feedback signal. For example, the feedback coefficient can be:

[0053] sensor_weight=0.5+0.1*math.sin

[0054] wherein the feedback coefficient is sensor_weight, and math.sin is the sine function value.

[0055] In this embodiment, the environmental dynamic parameter is used to reflect the change frequency of the environment where the embodied intelligent device is located and the complexity of the environment, and the like. The environmental dynamic parameter can be obtained by analyzing and processing the trial collection data, so as to better reflect the actual environment. For example, the trial collection data can include light intensity values and temperature values, and the change frequency of the environment can be analyzed based on the light intensity values and the temperature values, and the change frequency is taken as the environmental dynamic parameter. The collection task complexity parameter can be determined according to the parameter requirements in the trial collection parameter set. For example, the higher the required accuracy, the greater the collection task complexity parameter; otherwise, the smaller the collection task complexity parameter.

[0056] Specifically, the environmental dynamic parameter and the collection task complexity parameter can be weighted and combined to obtain an environment quantization index. Through the environment quantization index, the influence of the environment and the collection task on the sensor configuration can be comprehensively reflected. For example, the environmental dynamic parameter and the collection task complexity parameter are normalized respectively, and the value range after the normalization processing can be [0, 1]. The weight of the environmental dynamic parameter is set to 0.6, and the weight corresponding to the collection task complexity parameter is set to 0.4, and the weighted sum is performed, and the value obtained after the weighted sum is taken as the environment quantization index.

[0057] Optionally, the environmental quantification index and the collection task difficulty index are fused based on the feedback coefficient to obtain a calculation formula of the sensor configuration intensity value as follows:

[0058] sensor_config = (sensor_weight * task_env.difficulty + (1-sensor_weight) * env_factor)

[0059] wherein, sensor_config is the sensor configuration intensity value, sensor_weight is the feedback coefficient, env_factor is the environmental quantification index, and task_env.difficulty is the collection task difficulty index. It should be noted that the collection task difficulty index can be determined based on at least one of the accuracy requirement of the trial collection data, the number of sensors included in the embodied intelligent device, and the number of obstacles in the environment in which the embodied intelligent device is located.

[0060] In a specific implementation, at least one of the type of sensor, the sampling frequency, and the spatial resolution can be determined by the sensor configuration intensity value. Specifically, the higher the sensor configuration intensity value, the higher the accuracy of the sensor required; otherwise, the lower the accuracy of the sensor required. Based on the sensor configuration intensity value, the accuracy of the sensor can be determined, and based on the accuracy of the sensor, the type of sensor can be selected. For the sampling frequency of the sensor, the greater the sensor configuration intensity value, the higher the sampling frequency of the sensor required to collect more delicate data changes; otherwise, the lower the sampling frequency of the sensor required. For the spatial resolution of the required sensor, the greater the sensor configuration intensity value, the higher the spatial resolution of the sensor required to collect more abundant spatial information; otherwise, the lower the spatial resolution of the sensor required. The type of sensor, the sampling frequency, and the spatial resolution determined above, as well as other data collection requirements, collectively constitute the data collection specification.

[0061] In the present embodiment, the sampling frequency can also be determined by the action space dimension of the embodied intelligent device. Specifically, the action space dimension can be the degree of freedom of the embodied intelligent device that can perform actions, for example, the joint flexibility of a robotic arm. The formula for determining the sampling frequency is as follows:

[0062] sample_rate = 30 * (1 + math.log(action_dim + 1))

[0063] wherein, sample_rate is a sampling frequency, action_dim is a dimension of an action space, and math.log is a natural logarithm function. It should be noted that, in order to ensure that the sampling frequency does not exceed the maximum value supported by hardware, the determined sampling frequency needs to be hardware-limited. For example, the upper limit of the sampling frequency can be set to 100 Hz. Further, among a plurality of hardware-supported frequency discrete values set in advance, a frequency discrete value with the smallest difference from the sampling frequency can be determined, and the determined frequency discrete value is updated as the sampling frequency. For example, the sampling frequency is 71 Hz, and the hardware only supports positive multiples of 10 Hz, so the frequency discrete value with the smallest difference from the sampling frequency is 70 Hz, and the sampling frequency is updated to 70 Hz to ensure that it is supported by hardware.

[0064] In some embodiments, the sampling frequency can be determined by the dimension of the action space of the embodied intelligent device, or the sampling frequency can be determined by the calculated sensor configuration intensity value. Alternatively, after using the two methods to calculate the sampling frequency, mutual correction is performed to obtain the final sampling frequency.

[0065] Optionally, a corresponding relationship between different sensor configuration intensity values and collection parameters can be set in advance, based on the corresponding relationship, a current collection parameter corresponding to the currently determined sensor configuration intensity value is determined, and based on the current collection parameter, a data collection specification is generated. Exemplarily, the collection parameter can be the type of the sensor, such as a high-precision type and a low-precision type.

[0066] In the embodiment, when determining the data collection specification, the quality of the trial collection data, the change frequency of the environment, the complexity of the environment, and the complexity of the collection task are fully considered, so that the obtained data collection specification can better adapt to the actual environment of the embodied intelligent device and the collection task requirements, and the accuracy and effectiveness of the delivered data are improved.

[0067] In some embodiments, the data processing method of the present application further comprises: when performing the mass production collection operation, at least one of the following monitoring is further performed: monitoring whether the point cloud density collected by the laser radar of the embodied intelligent device is greater than a target density; monitoring whether the data collection interval of the inertial measurement unit of the embodied intelligent device is less than a target time; monitoring whether the action sequence performed by the embodied intelligent device is complete.

[0068] In the mass production collection process, in order to ensure the integrity and accuracy of the data collected in the mass production, the collected data is monitored in real time, so that the abnormalities generated in the mass production process can be found in time.

[0069] Specifically, to ensure the accuracy of the data collected by the laser radar of the embodied smart device, the point cloud density collected by each laser radar on the embodied smart device can be determined. If the point cloud density collected by each laser radar is greater than the target density, it indicates that the collected point cloud data is accurate. Otherwise, it indicates that there is an abnormality in the laser radar collection process. The target density can be determined based on the model and performance of the laser radar. For example, the target density can be 1600 points per square meter.

[0070] The inertial measurement unit is an important sensor for measuring the motion state of the embodied smart device, and the size of the data collection interval directly affects the real-time and accuracy of the data. To ensure the real-time of the data, it is necessary to monitor whether the data collection interval of the inertial measurement unit is less than the preset target time. The target time can be determined based on the application scenario of the embodied smart device, for example, in the scenarios of fast motion of robots and obstacle avoidance of drones, the target time can be set to be relatively short, such as 10 milliseconds. For slow moving scenarios, the target time can be set to be relatively long, such as 50 milliseconds. If the data collection time interval is less than the preset target time, it indicates that the real-time of the inertial measurement unit meets the actual demand; otherwise, it indicates that the real-time of the inertial measurement unit does not meet the actual demand.

[0071] The action sequence is composed of a series of operation actions, and is important information reflecting the motion state of the embodied smart device. To ensure the integrity of the data collected in mass production, it is necessary to monitor whether the action sequence executed by the embodied smart device matches the preset action sequence. If they match, it indicates that the action sequence is complete; otherwise, it indicates that the action sequence is incomplete. Alternatively, the matching of the action sequence and the preset action sequence includes that each action in the action sequence is the same as the action contained in the preset action sequence, and the order of each action in the action sequence is consistent with the order of the action contained in the preset action sequence. Illustratively, the similarity difference value between the action sequence and the preset action sequence can be determined by the DTW (Dynamic Time Warping) algorithm. If the similarity difference value is less than the preset difference value, it indicates that the action sequence matches the preset sequence; otherwise, it indicates that they do not match. Illustratively, the preset difference value can be set to 0.15.

[0072] In specific implementation, if the point cloud density is greater than the target density, the data collection interval is less than the target time, and the action sequence is complete, it can be determined that the mass production collection operation is correct, and the mass production collection can continue. If any of the above conditions is not met, the collected mass production data needs to be deleted, and the mass production collection operation needs to be performed again to ensure the accuracy and integrity of the collected data.

[0073] In this embodiment, by monitoring the point cloud density, data acquisition interval and action sequence, problems in the production acquisition process can be discovered and corrected in a timely manner, ensuring the accuracy and integrity of the target acquisition data.

[0074] In order to save storage space, an optional embodiment of the present application provides for packaging target acquisition data, thereby reducing the amount of stored data, which is implemented as follows.

[0075] Optionally, after performing the production acquisition operation based on the data acquisition specification to obtain the target acquisition data of the embodied intelligent device, the method further includes performing a packaging operation on the target acquisition data. Specifically, the packaging operation includes at least one of the following: converting the target acquisition data into a standard format to reduce the adaptation cost between the processing algorithms corresponding to the target acquisition data; compressing the target acquisition data to reduce the data amount; generating metadata corresponding to the target acquisition data, the metadata including scene parameters and / or sensor calibration parameters.

[0076] In a specific implementation, the target acquisition data can be converted into a standard format through a pre-constructed ROS (Robot Operating System) standard format converter, which facilitates the analysis and processing of the target acquisition data and the adaptation with the processing algorithms, thereby reducing the adaptation cost.

[0077] In addition, to reduce the data amount and the occupation of storage space, the target acquisition data can be compressed. Optionally, a hybrid coding technology of HEVC (High Efficiency Video Coding) video coding standard and ZSTD data compression algorithm can be used to compress the target acquisition data. The compression ratio is greater than or equal to 10:1.

[0078] Further, in the packaging process, metadata corresponding to the target acquisition data can also be generated. The metadata is data used to describe the target acquisition data, which can include scene parameters and / or sensor calibration parameters. The scene parameters are used to describe the scene information corresponding to the target acquisition data, such as lighting conditions, temperature and humidity, etc.; the sensor calibration parameters are used to describe the calibration information of the sensor, such as the position, pose and accuracy of the sensor, etc. In a specific implementation, XML (eXtensible Markup Language) and JSON (JavaScript Object Notation) metadata in two formats can be generated at the same time to meet the data format requirements of different systems or applications, thereby improving the compatibility of the data. The metadata conforms to the ISO / IEC 23000-12 standard.

[0079] The embodiment can convert the target collection data into a standard format in the process of packaging the target collection data, reduce the adaptation cost between processing algorithms, and improve the data processing efficiency; at the same time, the target collection data is compressed to reduce the data amount and save the storage space; and the metadata corresponding to the target collection data is generated to facilitate the analysis and application of the target collection data in the future, which is beneficial to improve the efficiency and usability of data delivery.

[0080] In order to ensure the security of the target collection data, the application embodiment provides an optional embodiment for privacy protection processing of the target collection data, thereby reducing the risk of data leakage, and the specific implementation is as follows.

[0081] Optionally, after performing the mass production collection operation based on the data collection specification to obtain the target collection data of the embodied intelligent device, the method further includes: performing a privacy protection operation on the target collection data. The privacy protection operation includes at least one of the following: desensitizing the target collection data; generating key information bound to the target collection data to encrypt the target collection data and / or decrypt the target collection data; and recording access permissions of the target collection data by using a block chain to verify the access permissions of the target collection data.

[0082] In a specific implementation, the target collection data can be subjected to data recognition to identify different sensitive data in the target collection data. If the sensitive data is structured data, a first target processing strategy matched with the sensitive data is determined based on a sensitive level corresponding to a sensitive field in the structured data; if the sensitive data is image format unstructured data, a first target processing strategy matched with the sensitive data is determined based on a sensitive level corresponding to a sensitive area in the image; and if the sensitive data is text format unstructured data, a first target processing strategy matched with the sensitive data is determined based on a sensitive level corresponding to a sensitive event in the text.

[0083] The structured data refers to data with fixed format and explicit structure, for example, identity card number, work number and other structured data matched by using a regular expression library. In addition, the structured data also includes specific fields storing sensitive information. For example, the middle segment of the identity card number is a sensitive field.

[0084] The sensitive level is a classification of the sensitivity of the sensitive field in the structured data, which is used to indicate the importance and protection requirement of the data. The sensitive level is usually classified into multiple sensitive levels from low to high, such as a first sensitive level L1, a second sensitive level L2, a third sensitive level L3, a fourth sensitive level L4, and the like. The higher the level is, the more sensitive the data is, and more stringent protection measures need to be taken.

[0085] Specifically, the first target processing strategy includes at least one of full retention, partial masking, encrypted storage, and rejected storage. Full retention refers to storing or transmitting sensitive data in its original, unencrypted, and unmasked form without any modification or change. Partial masking is to hide or replace a part of the sensitive data, so that only part of the content is exposed when displayed or transmitted, and the remaining part of the content is masked. Encrypted storage is to encrypt the sensitive data by a specific encryption algorithm, convert the data into ciphertext form for storage, and only those who have the correct key or decryption algorithm can restore the original data from the ciphertext. Rejected storage refers to directly rejecting to store some highly sensitive data in the system, or rejecting to store data from a specific source or with specific characteristics.

[0086] For example, if the sensitive field is L1, the first target processing strategy is full retention; if the sensitive field is L2, the first target processing strategy is partial masking; if the sensitive field is L3, the first target processing strategy is encrypted storage; and if the sensitive field is the fourth sensitive level L4, the first target processing strategy is rejected storage.

[0087] Unstructured data in image format refers to data in image form without fixed format or structure, such as photos, pictures, scans, etc. It can be understood that the area with specific sensitive information in the image, for example, the sensitive area includes face, fingerprint, private space, etc. when the image data is collected by the robot with specific sensitive information.

[0088] In some embodiments, a modified YOLOv5s model is used to detect sensitive areas in image data.

[0089] Similarly, if the sensitive area is L1, the first target processing strategy is full retention; if the sensitive area is L2, the first target processing strategy is partial masking; if the sensitive area is L3, the first target processing strategy is encrypted storage; and if the sensitive area is L4, the first target processing strategy is rejected storage.

[0090] Unstructured data in text format refers to data in natural language text format without fixed format or structure, such as emails, documents, instant messages, etc. A sensitive event is a specific situation or information described or related in the text. For example, the private conversation between users collected by the robot in the home environment, which contains the home address, personal contact information, user's health condition, etc.

[0091] In an implementation, a RoBERTa fine-tuning model is used to identify sensitive events in the text.

[0092] Similarly, if the sensitive event is L1, the first target processing strategy is complete retention; if the sensitive event is L2, the first target processing strategy is partial masking; if the sensitive event is L3, the first target processing strategy is encrypted storage; if the sensitive event is L4, the first target processing strategy is denied storage.

[0093] Furthermore, to enhance data security, key information can be generated and bound to the target collected data. Specifically, quantum key distribution technology can be used to generate this key information. This key information can be used to encrypt and store the target collected data, ensuring data security during storage. Furthermore, when access to the target collected data is required, this key information can be used for decryption, ensuring data availability.

[0094] Furthermore, to track and record access to target data, blockchain technology can be used to record access rights to target data. Blockchain's decentralized and tamper-proof nature ensures the authenticity and reliability of access rights records. Whenever an access request is made, the access rights recorded in the blockchain are compared to determine whether the requester is eligible. If the requester's identity matches the access rights recorded in the blockchain, access to the target data is granted; otherwise, the access request is denied. This approach enables refined management of access to target data, further reducing the risk of data leaks.

[0095] This embodiment performs privacy protection processing on the target collected data to ensure that the user's privacy and sensitive information are effectively protected during the data delivery process, effectively improves the security of the data, reduces the risk of data leakage, is conducive to protecting data security, and further enhances the user's trust in data security.

[0096] In this embodiment, to facilitate data delivery, after performing mass production collection operations based on data collection specifications to obtain target collection data of the embodied smart device, it also includes: based on the data delivery method selected by the user, selecting at least one from application programming interface, file transfer and cloud storage to send the target collection data to the user.

[0097] Specifically, different data delivery methods can be provided on the user side for the user to choose. The data delivery methods include at least one of an application programming interface, file transfer and cloud storage. The user can choose multiple data delivery methods at the same time to ensure that different data usage requirements can be met. For example, if the user needs to perform data analysis during the idle time period, the user can choose the data delivery methods of file transfer and cloud storage. By choosing file transfer, the target acquisition data is sent to the storage location specified by the user for storage. By uploading the target acquisition data to the cloud server, cloud storage of the target acquisition data is realized, which facilitates data access and management through the Internet. For the scene that needs real-time data analysis, the application programming interface can be used to send the target acquisition data to the user to ensure the timeliness and availability of the data.

[0098] Further, after the target acquisition data is sent to the user according to the data delivery method, a notification of successful data delivery can be sent to the user side to prompt the user to check.

[0099] The embodiment supports multiple data delivery methods, including but not limited to application programming interface, file transfer and cloud storage, etc. Through diversified delivery methods, the needs of different users' business systems can be met. The appropriate delivery method can be flexibly selected according to the actual situation of the user, which improves the availability and timeliness of the data and reduces the time cost and complexity of data delivery.

[0100] In order to ensure that the target acquisition data can be accurately obtained, an optional embodiment is provided, which proposes a real-time feedback adjustment mechanism for the test sampling parameter set. The test sampling parameter set corresponding to the test sampling data that fails the quality verification is updated to ensure that the test sampling data meets the quality requirements, and the specific implementation is as follows.

[0101] Optionally, the data processing method further includes: if the test sampling data fails the quality verification, sending a parameter adjustment suggestion to the user; in response to a confirmation operation of the user on the parameter adjustment suggestion, generating a new test sampling parameter set; based on the new test sampling parameter set, reacquiring the test sampling data of the embodied intelligent device, and performing quality verification on the reacquired test sampling data.

[0102] In order to more clearly understand the real-time feedback adjustment mechanism proposed in the present application, please refer to Figure 2 In Figure 2In the embodiment, the user terminal, the trial sampling system and the sensor array interact with each other to obtain trial sampling data that meets quality verification. Specifically, the user terminal inputs scene requirements through a display interface, the trial sampling system obtains the scene requirements, generates a set of trial sampling parameters based on the scene requirements, sends a collection instruction to the sensor array based on the set of trial sampling parameters, and receives trial sampling data returned by the sensor array corresponding to the set of trial sampling parameters. The trial sampling data is subjected to quality verification. If the verification is passed, the data is qualified, and a data collection specification can be generated based on the set of trial sampling parameters and sent to the user terminal. If the verification fails, the data is abnormal, the set of trial sampling parameters can be adjusted, and a parameter adjustment suggestion can be generated based on the adjusted set of trial sampling parameters and sent to the user terminal for confirmation. In response to the user's confirmation operation on the parameter adjustment suggestion, the adjusted set of trial sampling parameters can be used as a new set of trial sampling parameters. In response to the user's return operation on the parameter adjustment suggestion, the adjusted set of trial sampling parameters can be adjusted again, and the parameter adjustment suggestion can be updated based on the re-adjusted set of trial sampling parameters and sent to the user terminal until the user's confirmation operation is received. Based on the latest obtained set of trial sampling parameters, the trial sampling data of the embodied intelligent device is re-acquired, and the re-acquired trial sampling data is subjected to quality verification.

[0103] It should be noted that, in order to avoid infinite loop, the number of adjustments to the set of trial sampling parameters can be less than or equal to a preset number. For example, the preset number can be 3. If the number of adjustments to the set of trial sampling parameters is equal to the preset number, and the user terminal's confirmation operation is still not received, an exception information can be generated and sent to the operation and maintenance terminal to timely inform the operation and maintenance personnel to solve the fault and avoid affecting the data delivery progress.

[0104] The embodiment first verifies the quality of the trial sampling data, generates a parameter adjustment suggestion for user confirmation when the quality verification fails, realizes intelligent adjustment of the set of trial sampling parameters, and is conducive to determining a set of trial sampling parameters that meet the requirements. Moreover, by setting an upper limit on the number of adjustments to the set of trial sampling parameters, resource waste caused by infinite loop adjustment of the set of trial sampling parameters is avoided, and the efficiency and accuracy of data processing are further improved.

[0105] In order to be able to describe the data processing process from multiple angles, the following describes the data processing process in combination with a data processing system for executing the data processing method and an operation standard operation procedure fusion scheme.

[0106] Figure 3 The architecture schematic diagram of the data processing system provided by the embodiment is shown, and the data processing process is described from the system architecture of data processing. As shown in FIG. 1, the data processing system includes a user terminal 1, a trial sampling system 2 and a sensor array 3. Figure 3As shown, the data processing system includes a front-end interaction layer, a core processing layer, and a delivery control layer. The front-end interaction layer receives natural language requirement input, converts the natural language into parameters through a requirement digitization parser. And, through a parameter verification module, quality verification is performed on the trial collection data. The core processing layer includes a trial collection verification engine, an intelligent packaging engine, and a privacy computing sandbox. The trial collection verification engine includes a dynamic specification generator for dynamic feedback on trial collection parameters and generates data collection specifications through a specification generation algorithm to perform mass production collection operations based on the data collection specifications. The intelligent packaging engine can compress target collection data according to efficient video coding compression standards, and perform metadata binding processing. Through the privacy computing sandbox, real-time desensitization processing and quantum encryption processing are performed on the target collection data. The delivery control layer provides multiple data delivery modes through a delivery mode selector, for example, the data delivery modes can include application programming interface, file transfer to edge nodes, and cloud storage through object storage services.

[0107] Figure 4 As shown is a flowchart of an operation standard procedure fusion scheme provided by an embodiment of the present application; as Figure 4 As shown, in the operation standard procedure fusion scheme, there are a trial collection verification stage, a mass production collection stage, and a packaging delivery stage. In the trial collection verification stage, the customer requirement is first input, i.e., requirement definition is performed. For example, the requirement can be "warehouse picking robot training data", based on the input requirement, a scene model is built, the actual running environment and workflow are simulated through the model, and then a trial collection parameter set, i.e., a parameter matrix, is obtained. Based on the trial collection parameter set, trial collection operation is performed, i.e., the trial collection verification process is entered. For example, 7-day trial collection can be performed, and the parameters in the trial collection parameter set can be "dynamically adjusting 3D camera frame rate from 30Hz to 60Hz" and "mechanical arm trajectory sampling accuracy ±5mm to ±2mm". Then, the quality of the trial collection data is verified, if it does not meet the standard, scene modeling is performed again, i.e., a new parameter matrix is obtained, and trial collection operation is performed again, and the trial collection data is verified again. If it meets the standard, a data collection specification is generated, i.e., a protocol is signed.

[0108] In the mass production collection stage, data mass production collection operation is performed through the data collection specification. The data obtained by mass production collection is subjected to heterogeneous data fusion to obtain target collection data. The target collection data is subjected to intelligent processing. The intelligent processing includes converting the target collection data into a standard format through a format standardization engine, and compressing the target collection data through a layered compression algorithm. Further, the target collection data can be subjected to multi-dimensional acceptance to verify the compliance and integrity of the target collection data. If the acceptance passes, it enters the packaging delivery stage. If the acceptance does not pass, the algorithm used in the intelligent processing can be adjusted, and the intelligent processing and multi-dimensional acceptance are performed again based on the adjusted algorithm until the acceptance passes.

[0109] In the encapsulation delivery stage, the target acquisition data is processed for secure delivery, for example, a digital signature verification file and a metadata description file are added to the target acquisition data, and the target acquisition data is transmitted according to an encrypted transmission protocol. Moreover, based on user feedback on the target acquisition data, demand iteration optimization is performed to perfect and continuously optimize the data processing system through closed-loop feedback.

[0110] It can be seen that the present scheme adopts an elastic data production module of 'trial production first and mass production later', constructs a trial production verification standard system, and proposes a method of dynamically generating data delivery specifications. The system automatically generates data delivery specifications according to user requirements, industry standards and data characteristics. Users can adjust data acquisition parameters in real time during the trial production stage, and the system will automatically update the delivery specifications, solving the problem of lack of unified data delivery specifications, making the data delivery process more flexible and efficient, and adapting to the needs of different users.

[0111] The data processing method embodiments of the present application are described in detail above, and the data processing device embodiments of the present application are described in detail below. It should be understood that the description of the data processing method embodiments and the description of the data processing device embodiments correspond to each other, and therefore, the parts not described in detail can be referred to the previous method embodiments. Figures 1 to 4 Figure 5 The data processing method embodiments of the present application are described in detail above, and the data processing device embodiments of the present application are described in detail below. It should be understood that the description of the data processing method embodiments and the description of the data processing device embodiments correspond to each other, and therefore, the parts not described in detail can be referred to the previous method embodiments.

[0112] Figure 5 Fig. 1 shows a structure schematic diagram of a data processing device provided by an embodiment of the present application. As shown in Fig. 1, the data processing device provided by the embodiment of the present application comprises: Figure 5 A parameter set determination module 510 is configured to determine a trial production parameter set based on data acquisition requirements of a user;

[0113] A trial production operation module 520 is configured to perform a trial production operation based on the trial production parameter set to obtain trial production data of the embodied intelligent device;

[0114] A specification generation module 530 is configured to generate a data acquisition specification based on the trial production parameter set if the trial production data passes quality verification;

[0115] A mass production acquisition module 540 is configured to perform a mass production acquisition operation based on the data acquisition specification to obtain target acquisition data of the embodied intelligent device.

[0116]

[0117] ​​In an embodiment of the present application, the specification generation module 530 is further configured to generate a feedback coefficient based on the trial parameter set; calculate an environment quantization index based on the environment dynamicity parameter and the collection task complexity parameter; fuse the environment quantization index and the collection task difficulty index based on the feedback coefficient to obtain a sensor configuration intensity value, the sensor configuration intensity value being used to represent at least one of the type of required sensor, the sampling frequency, and the spatial resolution; and generate the data collection specification based on the sensor configuration intensity value.

[0118] In an embodiment of the present application, the mass production collection operation includes at least one of the following: monitoring whether the point cloud density collected by the laser radar of the body-worn intelligent device is greater than a target density; monitoring whether the data collection interval of the inertial measurement unit of the body-worn intelligent device is less than a target time; and monitoring whether a sequence of actions performed by the body-worn intelligent device is complete.

[0119] In an embodiment of the present application, the apparatus further includes a packaging module configured to perform a packaging operation on the target collection data after performing the mass production collection operation based on the data collection specification to obtain the target collection data of the body-worn intelligent device; and wherein the packaging operation includes at least one of the following: converting the target collection data into a standard format to reduce the adaptation cost between processing algorithms corresponding to the target collection data; compressing the target collection data to reduce the data volume; and generating metadata corresponding to the target collection data, the metadata including scene parameters and / or sensor calibration parameters.

[0120] In an embodiment of the present application, the apparatus further includes a privacy protection module configured to perform a privacy protection operation on the target collection data after performing the mass production collection operation based on the data collection specification to obtain the target collection data of the body-worn intelligent device; and wherein the privacy protection operation includes at least one of the following: desensitizing the target collection data; generating key information bound to the target collection data to encrypt the target collection data and / or decrypt access to the target collection data; and recording access rights of the target collection data using a blockchain to verify the access rights of the target collection data.

[0121] In an embodiment of the present application, the apparatus further includes a data sending module configured to send the target collection data to the user based on at least one of an application programming interface, file transfer, and cloud storage according to a data delivery method selected by the user after performing the mass production collection operation based on the data collection specification to obtain the target collection data of the body-worn intelligent device.

[0122] In one embodiment of the present application, the device also includes a suggestion sending module for sending parameter adjustment suggestions to the user if the test data fails the quality verification; generating a new test parameter set in response to the user's confirmation operation on the parameter adjustment suggestion; based on the new test parameter set, reacquiring the test data of the embodied smart device, and performing quality verification on the reacquired test data.

[0123] Below, reference Figure 6 To describe the electronic device according to the embodiment of the present application. Figure 6 Shown is a schematic structural diagram of an electronic device provided by an exemplary embodiment of the present application.

[0124] like Figure 6 As shown, the electronic device 60 includes one or more processors 601 and a memory 602 .

[0125] The processor 601 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 60 to perform desired functions.

[0126] The memory 602 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 601 may run the program instructions to implement the data processing methods of the various embodiments of the present application described above and / or other desired functions. In one example, the electronic device 60 may further include: an input device 603 and an output device 604, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0127] The input device 603 may include, for example, a keyboard, a mouse, and the like.

[0128] The output device 604 can output various information to the outside, and can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto.

[0129] Of course, to simplify, Figure 6 Only some of the components related to the present application in the electronic device 60 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device 60 may further include any other appropriate components according to specific application scenarios.

[0130] In addition to the method and the device described above, the embodiments of the present application can also be a computer program product, which includes computer program instructions, which, when executed by a processor, cause the processor to perform the steps of the data processing method according to various embodiments of the present application described above in the specification.

[0131] The computer program product can be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, C++, etc., and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server.

[0132] In addition, the embodiments of the present application can also be a computer readable storage medium, which stores computer program instructions, which, when executed by a processor, cause the processor to perform the steps of the data processing method according to various embodiments of the present application described above in the specification.

[0133] The computer readable storage medium can employ any combination of one or more computer readable media. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium can include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any suitable combination of the above. More specific examples (a non-exhaustive list) of the computer readable storage medium include an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0134] The basic principles of the present application are described above in combination with specific embodiments, but it should be noted that the advantages, advantages, effects, etc. mentioned in the present application are only examples and are not limited, and these advantages, advantages, effects, etc. cannot be considered as the necessary possession of each embodiment of the present application. In addition, the above-mentioned specific details are only for the purpose of example and for the purpose of understanding, and the above-mentioned details do not limit the present application to the above-mentioned specific details.

[0135] The block diagrams of the devices, devices, equipment, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "include," "comprise," "have," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0136] It should also be noted that in the apparatus, device, and method of the present application, each component or each step can be decomposed and / or recombined, and such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0137] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0138] The above description has been provided for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A data processing method, characterized in that: include: Determine the test sampling parameter set based on the user's data collection requirements; Performing a test sampling operation based on the test sampling parameter set to obtain test sampling data of the embodied intelligent device; If the test mining data passes the quality verification, generating a data collection specification based on the test mining parameter set; A mass production collection operation is performed based on the data collection specification to obtain target collection data of the embodied intelligent device.

2. The data processing method according to claim 1, wherein: Generating a data acquisition specification based on the test acquisition parameter set includes: generating a feedback coefficient based on the test mining parameter set; Calculate the environmental quantitative index based on the environmental dynamics parameters and the acquisition task complexity parameters; The environmental quantification index and the acquisition task difficulty index are integrated based on the feedback coefficient to obtain a sensor configuration strength value, where the sensor configuration strength value is used to represent at least one indicator of the required sensor type, sampling frequency, and spatial resolution; The data collection specification is generated based on the sensor configuration strength value.

3. The data processing method according to claim 1, wherein: Also includes: When performing the mass production acquisition operation, at least one of the following monitoring items is also performed: monitoring whether the point cloud density collected by the laser radar of the embodied intelligent device is greater than the target density; monitoring whether a data collection interval of an inertial measurement unit of the embodied intelligent device is less than a target time; Monitor whether the action sequence executed by the embodied intelligent device is complete.

4. The data processing method according to any one of claims 1 to 3, characterized in that: After performing the mass production collection operation based on the data collection specification to obtain the target collection data of the embodied smart device, the method further includes: Performing a packaging operation on the target collected data; The encapsulation operation includes at least one of the following: Converting the target collected data into a standard format to reduce the adaptation cost between processing algorithms corresponding to the target collected data; Compressing the target collected data to reduce the amount of data; Metadata corresponding to the target acquisition data is generated, where the metadata includes scene parameters and / or sensor calibration parameters.

5. The data processing method according to any one of claims 1 to 3, characterized in that: After performing the mass production collection operation based on the data collection specification to obtain the target collection data of the embodied smart device, the method further includes: Performing a privacy protection operation on the target collected data; The privacy protection operation includes at least one of the following: Desensitizing the target collected data; Generate key information bound to the target collected data so as to encrypt and store the target collected data and / or decrypt and access the target collected data; The blockchain is used to record the access rights of the target collected data so as to verify the access rights of the target collected data.

6. The data processing method according to any one of claims 1 to 3, characterized in that: After performing the mass production collection operation based on the data collection specification to obtain the target collection data of the embodied smart device, the method further includes: Based on the data delivery method selected by the user, at least one of application programming interface, file transfer and cloud storage is selected to send the target collected data to the user.

7. The data processing method according to any one of claims 1 to 3, characterized in that: Also includes: If the test data fails the quality verification, a parameter adjustment suggestion is sent to the user; In response to the user's confirmation operation on the parameter adjustment suggestion, generating a new test mining parameter set; Based on the new test sampling parameter set, the test sampling data of the embodied intelligent device is reacquired, and quality verification is performed on the reacquired test sampling data.

8. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and the computer program is used to execute the data processing method according to any one of claims 1 to 7.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is used to execute the data processing method described in any one of claims 1 to 7.

10. A computer program product, characterized in that The computer program product includes instructions, which, when executed on an electronic device, enable the electronic device to implement the data processing method according to any one of claims 1 to 7.

Images