Data labeling method, device, equipment and computer storage medium

Abstract

The embodiment of the application provides a kind of data labeling method, device, equipment and computer storage medium, the data labeling method includes obtaining the first data of target object that first roadside device acquisition device collects;First data based on target object is labeled to target object, and first labeling result is obtained;In response to the calibration input of first labeling result, second labeling result after calibration processing of first labeling result is obtained;Second labeling result is sent to the preset cloud server, for the preset cloud server to send second labeling result to second roadside device, for when second roadside device obtains the second data of target object, based on second labeling result and second data, target object is labeled.According to the embodiment of the application, a variety of data can be synchronously labeled, the accuracy of data labeling result when object detection is improved, and the utilization of artificial labeling data can be improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of machine learning, and particularly relates to a data labeling method and device, equipment and a computer storage medium. BACKGROUND

[0002] With the development of intelligent transportation and Internet technology, automatic labeling of traffic road images has become a current research hotspot, and in particular, the technology of automatically labeling traffic participants such as people, vehicles and motorcycles on traffic roads has a profound impact on automatic driving. Therefore, how to accurately label traffic road images has become a problem to be solved.

[0003] In the prior art, object detection is mainly performed through deep learning to preliminarily realize automatic labeling of data, and subsequent manual review is performed to ensure the quality of the labeled data. However, the above method relies heavily on the object recognition accuracy of the deep learning framework in the early stage, and a large amount of manual work is required to ensure the quality of the labeled data in the later stage, which has the problems of low accuracy of data labeling results and low utilization rate of manual work. SUMMARY

[0004] The embodiments of the present application provide a data labeling method, device, equipment and computer storage medium, which can solve the problems of low accuracy of data labeling results and low utilization rate of manual work in the prior art.

[0005] In a first aspect, the embodiments of the present application provide a data labeling method, which comprises:

[0006] obtaining first data of a target object collected by a collection device in a first roadside device;

[0007] labeling the target object based on the first data of the target object to obtain a first labeling result;

[0008] in response to a calibration input of the first labeling result, obtaining a second labeling result after calibration processing of the first labeling result;

[0009] sending the second labeling result to a preset cloud server, so that the preset cloud server sends the second labeling result to a second roadside device, and the second roadside device labels the target object based on the second labeling result and second data of the target object when the second data is obtained.

[0010] In an optional embodiment, when the collection device is multiple, the obtaining of the first data of the target object collected by the collection device in the first roadside device comprises:

[0011] temporally and spatially synchronously calibrating the multiple collection devices;

[0012] The plurality of acquisition devices calibrated by the space-time synchronization acquire first data of the target object in a preset observation range.

[0013] The first data of the target object is acquired.

[0014] In an optional implementation, the first data of the target object is based on the first data of the target object to label the target object, and a first labeling result is obtained, comprising:

[0015] According to a preset fusion perception algorithm, the first data of the target object acquired by the plurality of acquisition devices is fused;

[0016] The first data of the target object is based on the first data of the target object to label the target object, and a first labeling result is obtained.

[0017] In an optional implementation, the second labeling result includes a control instruction of the target object; the second labeling result is sent to the preset cloud server, so that the preset cloud server sends the second labeling result to a second road side device, and when the second road side device acquires second data of the target object, the target object is labeled based on the second labeling result and the second data, comprising:

[0018] The second labeling result is sent to the preset cloud server, so that the preset cloud server determines a target road side device of a driving section corresponding to the control instruction according to the control instruction of the target object, and sends the second labeling result to the target road side device, so that when the target road side device acquires second data of the target object, the target object is labeled based on the second labeling result and the second data;

[0019] The second road side device includes the target road side device.

[0020] In an optional implementation, the acquisition device includes a camera, a laser radar, and a millimeter wave radar.

[0021] In a second aspect, the embodiments of the present application provide a data labeling device, comprising:

[0022] The acquisition module is configured to acquire first data of a target object collected by an acquisition device in a first road side device.

[0023] The labeling module is configured to label the target object based on the first data of the target object, and obtain a first labeling result.

[0024] a response module, configured to obtain a second annotation result after calibration processing of the first annotation result in response to a calibration input of the first annotation result;

[0025] a sending module, configured to send the second annotation result to a preset cloud server, so that the preset cloud server sends the second annotation result to a second road-side device, and so that the second road-side device performs annotation on the target object based on the second annotation result and second data of the target object when the second data is acquired.

[0026] In an optional implementation, the apparatus further includes a calibration module and a collection module.

[0027] The calibration module is configured to perform spatio-temporal synchronization calibration on the plurality of collection apparatuses.

[0028] The collection module is configured to collect first data of the target object in a preset observation range by using the plurality of collection apparatuses that have been subjected to spatio-temporal synchronization calibration.

[0029] The acquisition module is further configured to acquire the first data of the target object.

[0030] In an optional implementation, the data annotation apparatus further includes a fusion module.

[0031] The fusion module is configured to fuse the first data of the target object collected by the plurality of collection apparatuses according to a preset fusion perception algorithm.

[0032] The annotation module is further configured to perform annotation on the target object based on the fused first data of the target object, to obtain the first annotation result.

[0033] In an optional implementation, the sending module is further configured to send the second annotation result to the preset cloud server, so that the preset cloud server determines a target road-side device of a driving section corresponding to a control instruction of the target object according to the control instruction of the target object, sends the second annotation result to the target road-side device, and so that the target road-side device performs annotation on the target object based on the second annotation result and second data of the target object when the second data is acquired.

[0034] The second road-side device includes the target road-side device.

[0035] In an optional implementation, the collection apparatus includes a camera, a laser radar, and a millimeter wave radar.

[0036] In a third aspect, an electronic device is provided, which includes a processor and a memory storing computer program instructions.

[0037] The processor implements the data labeling method as described in any one of the embodiments of the first aspect when executing the computer program instructions.

[0038] In a fourth aspect, an embodiment of the present application provides a computer storage medium, and the computer storage medium stores computer program instructions. The computer program instructions are executed by a processor to implement the data labeling method as described in any one of the embodiments of the first aspect.

[0039] In a fifth aspect, an embodiment of the present application provides a computer program product. Instructions in the computer program product are executed by a processor of an electronic device to cause the electronic device to perform the data labeling method as described in any one of the embodiments of the first aspect.

[0040] The data labeling method, device, equipment and computer storage medium provided by the embodiments of the present application can obtain first data of a target object collected by a collection device in a first roadside device, label the target object based on the first data of the target object, and obtain a first labeling result. The single roadside device can fuse multiple sensor data, and thus synchronize the labeling of the multiple data. Then, in response to a calibration input of the first labeling result, a second labeling result after calibration processing of the first labeling result is obtained, and the second labeling result is sent to a preset cloud server. The preset cloud server sends the second labeling result to a second roadside device, so that when the second roadside device obtains second data of the target object, the target object is labeled based on the second labeling result and the second data. In this way, the accuracy of the data labeling result in object detection can be improved, and the utilization rate of artificial labeling data can be improved through the edge cloud-based collaborative mechanism. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments of the present application. For those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0042] Figure 1 is a flowchart of a data labeling method provided by an embodiment of the present application;

[0043] Figure 2 is a typical intersection diagram based on an edge cloud provided by an embodiment of the present application;

[0044] Figure 3 is a structural diagram of a data labeling device provided by an embodiment of the present application;

[0045] Figure 4 is a structural diagram of an electronic device provided by an embodiment of the present application. Detailed Implementation

[0046] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0048] As described in the background section, existing technologies suffer from low accuracy in data annotation results and low utilization of manual labor. To address these issues, this application provides a data annotation method, apparatus, device, and computer storage medium. This data annotation method acquires first data of a target object collected by a data acquisition device in a first roadside device, annotates the target object based on this first data, obtains a first annotation result, and, in response to a calibration input for the first annotation result, obtains a second annotation result after calibration processing. This second annotation result is then sent to a preset cloud server, which in turn sends it to a second roadside device. Thus, when the second roadside device acquires the second data of the target object, it annotates the target object based on the second annotation result and the second data. This allows for simultaneous annotation of multiple data types through the fusion of data from multiple sensors in a single roadside device, improving the accuracy of data annotation results during object detection. Furthermore, the collaborative mechanism based on edge cloud technology enhances the utilization of manually annotated data. The data annotation method provided in this application is described below.

[0049] Figure 1 A flowchart illustrating a data annotation method provided in one embodiment of this application is shown.Figure 1 A flowchart of a data labeling method provided by an embodiment of the present application is shown.

[0050] As shown in the figure, the data labeling method can specifically include the following steps: Figure 1

[0051] S110, acquiring first data of a target object collected by a collection device in a first roadside device.

[0052] The first roadside device can be a device arranged at a road intersection or in the middle of a road section, for example, a road intersection roadside pole, and can be responsible for collecting data of the entire intersection. The collection device can be a device in the first roadside device for collecting roadside data, for example, a roadside camera, a laser radar, a millimeter wave radar, etc. The target object can be a main traffic participant in the road, for example, can include vehicles, pedestrians, bicycles, motorcycles, etc., and can also be an obstacle, a road surface marker, and a signal lamp, etc. The first data can be image, point cloud, millimeter wave radar, etc. data collected by the collection device.

[0053] S120, labeling the target object based on the first data of the target object to obtain a first labeling result.

[0054] In the embodiment of the present application, the target object is labeled based on the first data of the target object collected by the collection device to obtain a first labeling result, wherein the first labeling result can include but is not limited to the size, moving speed, control command, appearance feature, etc. of the target object.

[0055] S130, in response to a calibration input of the first labeling result, obtaining a second labeling result after calibration processing of the first labeling result.

[0056] The calibration input can be a correction input of the user to the first labeling result, which can label the objects not labeled in the first labeling result, or can correct the incorrect labeling result in the first labeling result. The second labeling result can be a labeling result after manual calibration.

[0057] S140, sending the second labeling result to a preset cloud server, so that the preset cloud server sends the second labeling result to a second roadside device, and when the second roadside device acquires second data of the target object, labels the target object based on the second labeling result and the second data.

[0058] ​The preset cloud server can be a cloud computing center providing real-time data processing and analysis decision, for example, can be an edge cloud. The second road side device can be a device arranged at a road side position of a road section, for example, can be a road section road side pole, and can be responsible for collection of data of the corresponding road section. After receiving the second labeling result sent by the preset cloud server, the second road side device labels the target object when collecting the second data of the target object, obtains a labeling result corresponding to the second data of the target object, calibrates the labeling result corresponding to the second data of the target object according to the second labeling result, and obtains a final labeling result.

[0059] In the embodiment of the application, the first data of the target object collected by the collection device in the first road side device is obtained, the target object is labeled based on the first data of the target object, and a first labeling result is obtained. The fusion of multiple sensor data of a single road side device can be performed, and thus multiple data can be synchronously labeled. Further, in response to the calibration input of the first labeling result, a second labeling result after calibration processing of the first labeling result is obtained, and the second labeling result is sent to the preset cloud server. The second labeling result is sent to the second road side device by the preset cloud server, so that the second road side device labels the target object based on the second labeling result and the second data of the target object when the second data of the target object is obtained. In this way, the accuracy of the data labeling result in object detection can be improved, and the utilization rate of artificial labeled data can be improved through the collaborative mechanism based on the edge cloud.

[0060] In some embodiments, when the collection device is multiple, the above S110: obtaining the first data of the target object collected by the collection device in the first road side device can specifically include:

[0061] Calibrating the multiple collection devices in time and space;

[0062] The multiple collection devices after time and space calibration collect the first data of the target object in the preset observation range;

[0063] Obtaining the first data of the target object.

[0064] The preset observation range can be an observation range corresponding to the collection device. Before collecting data, each collection device needs to be calibrated in time and space to enable the multiple collection devices to synchronously collect data.

[0065] In the embodiment of the application, the multiple collection devices are calibrated in time and space, so that the multiple collection devices can synchronously collect data. The multiple collection devices after time and space calibration collect the first data of the target object in the preset observation range, so that the obtained first data of the target object is synchronously collected data, and the accuracy of object detection and data labeling can be improved.

[0066] In some embodiments, the collection device includes a camera, a laser radar, and a millimeter wave radar.

[0067] In the embodiments of the present application, the plurality of collection devices can include a roadside camera, a laser radar, a millimeter wave radar, and a device capable of detecting objects. By collecting data through multiple different types of collection devices, the labeling error caused by labeling only one type of data can be avoided.

[0068] In some embodiments, the above S120: labeling the target object based on the first data of the target object to obtain a first labeling result, can specifically include:

[0069] According to a preset fusion perception algorithm, the first data of the target object collected by the plurality of collection devices is fused;

[0070] Based on the fused first data of the target object, the target object is labeled to obtain a first labeling result.

[0071] The preset fusion perception algorithm can be an algorithm capable of fusing data collected by different sensors. The preset fusion perception algorithm can use different deep learning frameworks, or can be based on existing labeled data for model training.

[0072] In one example, as shown in Figure 2 P0 is a roadside pole at an intersection, responsible for collecting data at the entire intersection. Based on the image, laser point cloud, and millimeter wave radar data collected by the collection device in the roadside pole P0, a fusion perception algorithm is used to detect the traffic participants, such as vehicles, in the fused data, thereby preliminarily completing the automatic labeling of the data.

[0073] In the embodiments of the present application, by fusing the first data of the target object collected by the plurality of collection devices according to the preset fusion perception algorithm, and labeling the target object based on the fused first data of the target object to obtain a first labeling result, multiple types of data can be labeled at the same time, the object detection accuracy is improved, and the accuracy of the data labeling result is further improved.

[0074] In some embodiments, the second labeling result can include a control instruction of the target object; the above S140 can specifically include:

[0075] The second labeling result is sent to a preset cloud server, so that the preset cloud server determines a target roadside device of a driving section corresponding to the control instruction according to the control instruction of the target object, and sends the second labeling result to the target roadside device, so that when the target roadside device obtains the second data of the target object, the target object is labeled based on the second labeling result and the second data; wherein the second roadside device includes the target roadside device.

[0076] Upon receiving the second annotation result, the preset cloud server determines the target object's driving direction based on the control commands within the second annotation result. Then, based on the target object's driving direction, it identifies the target roadside device on the corresponding road segment and sends the second annotation result to that target roadside device. Upon receiving the second annotation result, when the target object moves to the road segment corresponding to its driving direction, or when the target roadside device collects the target object's second data, it calibrates the annotation result corresponding to the target object's second data based on the distributed fusion perception strategy of the edge cloud, obtaining the final annotation result.

[0077] In one example, such as Figure 2 As shown, five roadside poles, P0, P1, P2, P3, and P4, are all uniformly connected to the edge cloud. The first annotation result based on the data collected from roadside pole P0 is manually calibrated to obtain the second annotation result, thus completing the detection and manual calibration of vehicles V1, V2, V3, and V4. The second annotation result is uploaded to the edge cloud. Under the coordination of the edge cloud, taking vehicle V3 as an example, regardless of the direction vehicle V3 travels, its control commands are recorded by the edge cloud. The second annotation result is then pushed to the roadside pole corresponding to the direction of vehicle V3's control command. For example, if vehicle V3 is traveling straight, its second annotation result will be pushed to roadside pole P4 by the edge cloud. Then, based on the edge cloud's distributed fusion perception strategy, roadside pole P4 will use this prior information of the second annotation result to calibrate the annotation results obtained in the automatic object detection and annotation stages, obtaining the final annotation result. Similarly, if vehicle V3 turns right, its second labeling result will be pushed to roadside pole P2; if vehicle V3 turns left, its second labeling result will be pushed to roadside pole P1; if vehicle V3 makes a U-turn, its second labeling result will be pushed to roadside pole P3.

[0078] In this embodiment, by sending the second annotation result to a preset cloud server, the manually calibrated second annotation result can be sent to the preset cloud server for unified management, improving the utilization rate of manual work. The preset cloud server determines the target roadside equipment corresponding to the control command of the target object on the corresponding driving segment, thereby accurately pushing data to the target roadside equipment and ensuring efficient data transmission. The preset cloud server sends the second annotation result to the target roadside equipment, so that when the target roadside equipment obtains the second data of the target object, it can annotate the target object based on the second annotation result and the second data, effectively improving the efficiency and accuracy of object detection and data annotation.

[0079] Figure 3 This is a schematic diagram of the structure of a data annotation device 300 according to an exemplary embodiment.

[0080] As Figure 3 shown, the data labeling apparatus 300 can include:

[0081] The acquisition module 301 is configured to acquire first data of a target object collected by a collection device in a first roadside device.

[0082] The labeling module 302 is configured to label the target object based on the first data of the target object to obtain a first labeling result.

[0083] The response module 303 is configured to obtain a second labeling result after calibration processing of the first labeling result in response to a calibration input of the first labeling result.

[0084] The sending module 304 is configured to send the second labeling result to a preset cloud server, so that the preset cloud server sends the second labeling result to a second roadside device, and the second roadside device labels the target object based on the second labeling result and second data of the target object when the second data of the target object is acquired.

[0085] In the embodiments of the present application, by acquiring the first data of the target object collected by the collection device in the first roadside device, labeling the target object based on the first data of the target object to obtain the first labeling result, the fusion of the multi-sensor data of a single roadside device can be performed, thereby synchronously labeling the multi-data. Further, in response to the calibration input of the first labeling result, the second labeling result after calibration processing of the first labeling result is obtained, the second labeling result is sent to the preset cloud server, so that the preset cloud server sends the second labeling result to the second roadside device, thereby the second roadside device labels the target object based on the second labeling result and the second data of the target object when the second data of the target object is acquired. In this way, the accuracy of the data labeling result in object detection can be improved, and the utilization rate of the artificial labeling data can be improved through the collaborative mechanism based on the edge cloud.

[0086] As an implementation manner of the present application, in order to enable the plurality of collection devices to synchronously collect data, the above-mentioned data labeling apparatus 300 can further include a calibration module and a collection module.

[0087] The calibration module is configured to perform time-space synchronization calibration on the plurality of collection devices.

[0088] The collection module is configured to collect the first data of the target object in a preset observation range by using the plurality of collection devices after time-space synchronization calibration.

[0089] The acquisition module 301 is further configured to acquire the first data of the target object.

[0090] As an implementation form of the present application, in order to simultaneously label multiple types of data, improve object detection accuracy, and further improve the accuracy of data labeling results, the data labeling device 300 can further include a fusion module.

[0091] The fusion module is configured to fuse the first data of the target object collected by the plurality of acquisition devices according to a preset fusion perception algorithm.

[0092] The labeling module 302 is further configured to label the target object based on the fused first data of the target object to obtain a first labeling result.

[0093] As an implementation form of the present application, in order to improve the utilization rate of manual work, accurately push data to the target road side device, ensure the efficiency of data transmission, and further improve the efficiency and accuracy of object detection and data labeling, the sending module 304 is further configured to send the second labeling result to a preset cloud server, so that the preset cloud server determines the target road side device of the driving section corresponding to the control instruction of the target object according to the control instruction, and sends the second labeling result to the target road side device, so that when the target road side device obtains the second data of the target object, the target road side device labels the target object based on the second labeling result and the second data.

[0094] The second road side device includes the target road side device.

[0095] As an implementation form of the present application, in order to collect data through a plurality of different types of acquisition devices, and avoid labeling errors caused by separate labeling of only one type of data, the acquisition device can include a camera, a laser radar, and a millimeter wave radar.

[0096] Figure 4 A hardware structure schematic diagram of an electronic device provided by an embodiment of the present application is shown.

[0097] The electronic device can include a processor 401 and a memory 402 having computer program instructions stored therein.

[0098] Specifically, the processor 401 can include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or can be configured to implement one or more integrated circuits of the embodiments of the present application.

[0099] The memory 402 can include mass storage for data or instructions. As an example and not by way of limitation, the memory 402 can include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disc (e.g., a compact disc (CD) or a digital versatile disc (DVD)), a solid-state drive (SSD), a USB drive, or a combination of two or more of these. Where appropriate, the memory 402 can include removable or non-removable (or fixed) media. Where appropriate, the memory 402 can be internal or external to the integrated gateway disaster recovery appliance. In particular embodiments, the memory 402 is non-volatile, solid-state memory.

[0100] The memory can include read-only memory (ROM), random-access memory (RAM), magnetic disk storage mediums, optical storage mediums, flash memory devices, electrical, optical, or other physical / tangible memory storage devices. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software that, when executed (by one or more processors), is operable to access the data and perform operations described with respect to the methods according to an aspect of the present disclosure.

[0101] The processor 401 implements any one of the data labeling methods in the above embodiments by reading and executing computer program instructions stored in the memory 402.

[0102] In one example, the electronic device can further include a communication interface 403 and a bus 410. As shown, the processor 401, the memory 402, and the communication interface 403 are connected through the bus 410 and complete communication among each other. Figure 4

[0103] The communication interface 403 is mainly used to realize the communication between the modules, devices, units and / or equipment in the embodiments of the present application.

[0104] ​Bus 410 includes a hardware, software, or both that couples components of data annotation device to each other. As an example and not by way of limitation, bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand (IB) interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or another suitable bus or a combination of two or more of these. Where appropriate, bus 410 can include one or more buses. Although this application describes and shows a particular bus, this application contemplates any suitable bus or interconnect.

[0105] The electronic device can execute the data annotation method in the embodiments of the present application based on the first data of the target object collected by the acquisition device in the first roadside device, thereby realizing the data annotation method in the embodiments of the present application in combination with Figure 1 The data annotation method described.

[0106] In addition, in combination with the data annotation method in the above embodiments, the embodiments of the present application can provide a computer storage medium to realize. The computer storage medium has computer program instructions stored thereon; the computer program instructions are executed by the processor to realize any one of the data annotation methods in the above embodiments.

[0107] It needs to be clear that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the present application.

[0108] The functions noted in the description of the structural block diagrams above can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, functional cards, and the like. When implemented in software, the elements of the present application are program or code segments that are used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. A "machine-readable medium" includes any medium that can store or transport information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and the like. The code segments can be downloaded via computer networks such as the Internet, intranets, and the like.

[0109] It is also important to note that the examples mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the steps mentioned in the examples, that is, the steps can be performed in the order mentioned in the examples, or in a different order from the examples, or several steps can be performed simultaneously.

[0110] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other processing devices to operate in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function / act specified in the flowchart and / or block diagram block or blocks. The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other processing devices to operate in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function / act specified in the flowchart and / or block diagram block or blocks.

[0111] The above merely describes a specific implementation of the present application. Those skilled in the art can clearly understand the specific working processes of the system, modules and units described above for the convenience and brevity of description, and can refer to the corresponding processes in the foregoing method embodiments, which will not be described herein again. It should be understood that the protection scope of the present application is not limited to this, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. A data annotation method, characterized in that, include: Acquire the first data of the target object collected by the acquisition device in the first roadside equipment; The target object is labeled based on the first data of the target object to obtain a first labeling result; In response to the calibration input of the first annotation result, a second annotation result after calibration processing of the first annotation result is obtained; The second annotation result is sent to a preset cloud server, so that the preset cloud server can send the second annotation result to a second roadside device, so that when the second roadside device obtains the second data of the target object, it can annotate the target object based on the second annotation result and the second data; The second annotation result includes control instructions for the target object; the step of sending the second annotation result to a preset cloud server, so that the preset cloud server can send the second annotation result to a second roadside device, so that when the second roadside device obtains the second data of the target object, it can annotate the target object based on the second annotation result and the second data, includes: The second annotation result is sent to the preset cloud server so that after receiving the second annotation result, the preset cloud server determines the driving direction of the target object according to the control command of the target object in the second annotation result, and determines the target roadside device of the driving segment corresponding to the driving direction according to the driving direction of the target object, and sends the second annotation result to the target roadside device so that the target roadside device can obtain the second data of the target object. When the target object moves to the driving segment corresponding to the driving direction, or when the target roadside device collects the second data of the target object, the annotation result corresponding to the second data of the target object is calibrated according to the second annotation result based on the distributed fusion perception strategy of edge cloud to obtain the final annotation result. The second roadside equipment includes the target roadside equipment.

2. The method according to claim 1, characterized in that, When there are multiple acquisition devices, acquiring the first data of the target object collected by the acquisition device in the first roadside equipment includes: Spatiotemporal synchronization calibration is performed on multiple of the aforementioned acquisition devices; Multiple acquisition devices, after being spatiotemporally synchronized and calibrated, acquire first data of the target object within a preset observation range; Obtain the first data of the target object.

3. The method according to claim 2, characterized in that, The step of annotating the target object based on the first data of the target object to obtain a first annotation result includes: The first data of the target object collected by multiple acquisition devices are fused according to a preset fusion perception algorithm; The target object is labeled based on the first data of the fused target object to obtain the first labeling result.

4. The method according to claim 2, characterized in that, The data acquisition device includes a camera, a lidar, and a millimeter-wave radar.

5. A data annotation device, characterized in that, The device includes: The acquisition module is used to acquire the first data of the target object collected by the acquisition device in the first roadside equipment; The annotation module is used to annotate the target object based on the first data of the target object, and obtain a first annotation result; The response module is used to obtain a second annotation result after calibration processing of the first annotation result in response to the calibration input of the first annotation result; The sending module is used to send the second annotation result to a preset cloud server, so that the preset cloud server can send the second annotation result to a second roadside device, so that when the second roadside device obtains the second data of the target object, it can annotate the target object based on the second annotation result and the second data; The second annotation result includes the control command of the target object; the sending module is further configured to send the second annotation result to the preset cloud server, so that after receiving the second annotation result, the preset cloud server determines the driving direction of the target object according to the control command of the target object in the second annotation result, and determines the target roadside device corresponding to the driving direction according to the driving direction of the target object, and sends the second annotation result to the target roadside device, so that the target roadside device can obtain the second data of the target object. When the target object moves to the driving section corresponding to the driving direction, or when the target roadside device collects the second data of the target object, the annotation result corresponding to the second data of the target object is calibrated according to the second annotation result based on the distributed fusion perception strategy of edge cloud to obtain the final annotation result; wherein the second roadside device includes the target roadside device.

6. The apparatus according to claim 5, characterized in that, The device also includes a calibration module and a data acquisition module; The calibration module is used to perform spatiotemporal synchronization calibration on multiple acquisition devices; The acquisition module is used to acquire first data of the target object within a preset observation range using multiple acquisition devices after spatiotemporal synchronization calibration. The acquisition module is also used to acquire the first data of the target object.

7. An electronic device, characterized in that, The device includes: a processor and a memory storing computer program instructions; the processor reads and executes the computer program instructions to implement the data annotation method as described in any one of claims 1-4.

8. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the data annotation method as described in any one of claims 1-4.

9. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device performs the data annotation method as described in any one of claims 1-4.

Images