Video data identification method and device, equipment and medium
By automatically identifying target image frames in video data with insufficient viewpoints, the high cost and inefficiency caused by manual frame-by-frame browsing in existing technologies are solved, achieving automation and high efficiency in video data recognition.
Patent Information
- Application Number
- CN202310602509.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-25
- Publication Date
- 2024-11-26
AI Technical Summary
Existing technologies require manual image browsing frame by frame to identify whether labeled video data meets usage requirements, resulting in high labor costs and low efficiency.
By acquiring video data that has been annotated, the system determines target image frames with a number of viewpoints less than or equal to a preset threshold based on the annotation data of each image frame, and judges whether the video data can be used for map drawing based on the number of target image frames, thereby reducing human intervention.
The elimination of the need for manual judgment on whether video data can be used for map creation reduces labor costs and improves recognition efficiency.
Smart Images

Figure CN119027847A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to a video data recognition method, device, equipment and medium. Background Art
[0002] Full Self Driving Map (FSD Map) can display accurate vehicle location information and rich road element data information. It is widely used in the field of autonomous driving. The main purpose of the Full Self Driving Map is to assist perception algorithms in autonomous driving systems to depict and annotate real-world road scenes. In the production process of the Full Self Driving Map, it is necessary to annotate each frame of the collected video data, and to identify and judge whether the annotated video data meets the use requirements, so as to improve the production precision and accuracy of the autonomous driving map.
[0003] In the prior art, when identifying and judging whether the labeled video data meets the usage requirements, it is necessary to manually browse the images in the labeled video data frame by frame, and further judge whether the labeled video data meets the usage requirements based on whether the labeled data of each frame of the image meets the usage requirements. Due to the huge amount of labeled data, there are problems of high labor cost and low efficiency. Summary of the invention
[0004] In order to solve the above technical problems, the present disclosure provides a video data identification method, device, equipment and medium.
[0005] A first aspect of an embodiment of the present disclosure provides a video data recognition method, the method comprising:
[0006] Obtain video data that has been annotated;
[0007] Determine at least one target image frame in the video data according to the annotation data of each image frame in the video data, where the target image frame is an image frame whose number of viewing angles included in the annotation data is less than or equal to a first preset threshold;
[0008] According to the number of target image frames, the video data is identified to determine whether the video data can be used for map drawing.
[0009] A second aspect of an embodiment of the present disclosure provides a video data recognition device, the device comprising:
[0010] A data acquisition module, used to acquire video data that has been annotated;
[0011] A first determination module is used to determine a target image frame in the video data according to the annotation data in the video data, where the target image frame is an image frame that lacks annotation data corresponding to a preset number of viewing angles;
[0012] The second determination module is used to identify the video data according to the number of target image frames, and determine whether the video data is target video data, where the target video data is video data that cannot be used for map drawing.
[0013] A third aspect of an embodiment of the present disclosure provides an electronic device, the device comprising:
[0014] Memory;
[0015] Processor; and
[0016] A computer program, wherein the computer program is stored in a memory and is configured to be executed by a processor to implement the video data recognition method of the first aspect as described above.
[0017] A fourth aspect of an embodiment of the present disclosure provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the video data recognition method as described in the first aspect above is implemented.
[0018] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:
[0019] The video data recognition method, apparatus, device and medium provided by the embodiments of the present disclosure can obtain video data that has been annotated with data, and determine at least one target image frame in the video data based on the annotation data of each image frame in the video data that has been annotated with data, wherein the target image frame is an image frame whose number of viewing angles included in the annotation data is less than or equal to a first preset threshold value, and then identify the video data based on the number of target image frames to determine whether the video data can be used for map drawing. Therefore, the target image frame can be determined by determining whether the number of viewing angles included in the annotation data of each image frame in the video data is less than or equal to a first preset threshold value, and then determine whether the video data can be used for map drawing based on the number of target image frames in the video data, without the need for manual judgment on whether the video data can be used for map drawing, thereby reducing labor costs and improving the recognition efficiency of video data. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0022] Figure 1 is a flow chart of a video data identification method provided by an embodiment of the present disclosure;
[0023] Figure 2 is a schematic diagram of an intersection with incomplete video data provided by an embodiment of the present disclosure;
[0024] Figure 3 is a flow chart of another video data identification method provided by an embodiment of the present disclosure;
[0025] Figure 4 is a flowchart of a method for determining a recompile operation provided by an embodiment of the present disclosure;
[0026] Figure 5 is a structural diagram of a video data recognition device provided by an embodiment of the present disclosure;
[0027] Figure 6 It is a structural schematic diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION
[0028] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.
[0029] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.
[0030] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.
[0031] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0032] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0033] Generally, when the prior art identifies and judges whether the annotated video data meets the usage requirements, it is necessary to manually browse the images in the annotated video data frame by frame, and further judge whether the annotated video data meets the usage requirements based on whether the annotated data of each frame meets the usage requirements. Due to the huge amount of annotated data, there are problems of high labor cost and low efficiency. To solve this problem, the embodiments of the present disclosure provide a video data identification method, which is introduced below in conjunction with specific embodiments.
[0034] Figure 1 It is a flowchart of a video data recognition method provided by an embodiment of the present disclosure. The method can be executed by a video data recognition device. The video data recognition device can be implemented in software and / or hardware. The video data recognition device can be configured in an electronic device, such as a server or a terminal or a server cluster, wherein the terminal can specifically include a mobile phone, a computer or a tablet computer, a vehicle-mounted terminal, or any device that can be used to process the video data recognition method.
[0035] like Figure 1 As shown, the video data identification method provided by the embodiment of the present disclosure includes the following steps.
[0036] S110: Obtain video data that has been annotated.
[0037] In an embodiment of the present disclosure, when the electronic device receives a video data recognition instruction, it obtains the video data that has been marked.
[0038] In the embodiments of the present disclosure, the video data that has been annotated can be understood as a video containing annotated data, and the video data that has been annotated is data obtained by annotating videos collected by cameras installed in different positions on the same vehicle, wherein images around the vehicle can be collected using cameras in different positions, and there are overlapping image collection ranges between the cameras. For example, cameras may be installed in front of the vehicle, in front left, in front right, behind, in rear left, and in rear right, respectively, and fisheye cameras with a relatively large collection range may be used in the front left, front right, rear left, and rear right.
[0039] Furthermore, the annotation data may include annotation attribute data used to determine the annotation information.
[0040] Exemplarily, the annotation information is a rectangular box of a specific color, such as a rectangular box with a color value of (255, 0, 0), and the annotation attribute data can be the value of the specific color, etc., which is not limited here.
[0041] Specifically, when the electronic device receives the video data identification instruction, it obtains the video data for data labeling from a preset database according to the video data identifier carried in the video data identification instruction.
[0042] S120. Determine at least one target image frame in the video data according to the annotation data of each image frame in the video data, where the target image frame is an image frame whose number of viewing angles included in the annotation data is less than or equal to a first preset threshold.
[0043] In an embodiment of the present disclosure, after acquiring video data that has been annotated, the electronic device determines at least one target image frame in the video data based on the annotated data of each image frame in the video data, where the target image frame is an image frame in which the number of viewing angles included in the annotated data is less than or equal to a first preset threshold.
[0044] In the embodiment of the present disclosure, the viewing angles include six viewing angles: a left front viewing angle, a front viewing angle, a right front viewing angle, a left rear viewing angle, a front rear viewing angle, and a right rear viewing angle.
[0045] The first preset threshold is a preset threshold used to determine whether an image frame is a target video frame.
[0046] Each video frame includes images of the surroundings of the vehicle collected at the same time and at the same position (e.g., images of six viewing angles), and the annotation data of each image frame may be annotation data obtained by annotating images of the surroundings of the vehicle collected at the same time and at the same position. Under normal circumstances, each image frame includes annotation data of six viewing angles. When the number of viewing angles in the annotation data of the image frame is less than or equal to a first preset threshold, the image frame is determined as a target image frame.
[0047] The target image frame can also be understood as an image frame with incomplete annotation data of the perspective.
[0048] Exemplarily, the first preset threshold is 2, and when the annotation data of the image frame lacks annotation data of two perspectives, the image frame is determined to be a target image frame.
[0049] S130. Identify the video data according to the number of target image frames, and determine whether the video data can be used for map drawing.
[0050] In the disclosed embodiment, after obtaining the target image frame, the electronic device identifies the video data according to the number of the target image frames and determines whether the video data can be used for map drawing.
[0051] Specifically, after obtaining the target image frame, the electronic device determines the number of target image frames in the video data, and then determines the integrity of the video data based on the number of target image frames, and then determines whether the video data can be used for map drawing based on the integrity.
[0052] Figure 2 is a schematic diagram of an intersection with incomplete video data provided by an embodiment of the present disclosure, such as Figure 2 The dashed boxes 001 and 002 are sections of video data that have not been annotated. At this intersection, since there is no annotated video data in the dashed boxes 001 and 002, the video data of the entire intersection cannot be used for map drawing.
[0053] In an embodiment of the present disclosure, it is possible to obtain video data that has been annotated with data, and determine at least one target image frame in the video data based on the annotation data of each image frame in the video data that has been annotated with data, wherein the target image frame is an image frame whose number of viewing angles included in the annotation data is less than or equal to a first preset threshold value. Then, based on the number of target image frames, the video data is identified to determine whether the video data can be used for map drawing. Thus, it is possible to determine the target image frame by determining whether the number of viewing angles included in the annotation data of each image frame in the video data is less than or equal to a first preset threshold value, and then determine whether the video data can be used for map drawing based on the number of target image frames in the video data. There is no need for manual judgment on whether the video data can be used for map drawing, thereby reducing labor costs and improving the recognition efficiency of video data.
[0054] Based on the above-mentioned embodiment of the present invention, S120 determines at least one target image frame in the video data according to the annotation data of each image frame in the video data, which can specifically include: for each image frame, determining the number of viewing angles included in the annotation data corresponding to each image frame; when the number of viewing angles is less than or equal to a first preset threshold, determining that the image frame is a target image frame.
[0055] Furthermore, in S130, the video data is identified according to the number of target image frames, and it is determined that the video data can be used for map drawing, which can specifically include: according to the road scene type corresponding to the video data and the number of target image frames, the video data is identified, and it is determined that the video data can be used for map drawing.
[0056] The road scene type is the type of road the vehicle is on, such as road, intersection, roundabout, etc.
[0057] Figure 3 is a flowchart of another video data identification method provided by an embodiment of the present disclosure. Figure 3 As shown, the video data identification method includes steps S310-S390, and the video data is identified according to the road scene type and the number of target image frames corresponding to the video data. Determining whether the video data can be used for map drawing may specifically include steps S340-S390.
[0058] S310: Obtain video data that has been annotated.
[0059] S320: Determine at least one target image frame in the video data according to the labeled data of each image frame in the video data.
[0060] In the embodiment of the present disclosure, the specific implementation of S310 - S320 is similar to that of S110 - S120 in the above embodiment, and will not be described in detail here.
[0061] S330: Determine the road scene type corresponding to the video data.
[0062] In an embodiment of the present disclosure, determining the road scene type corresponding to the video data may specifically include: determining the first road scene type corresponding to the last image frame in the video data as the road scene type corresponding to the video data; or, determining the second road scene type corresponding to each image frame in the video data from multiple second road scene types, and determining the second road scene type with the largest number among the multiple second road scene types as the road scene type corresponding to the video data.
[0063] In some examples, if the road scene type corresponding to the last image frame of the video data is that the vehicle is on the road, then the vehicle is on the road is determined as the road scene type corresponding to the video data.
[0064] In other examples, the road scene type corresponding to each image frame in the video data is determined. For example, if the video data contains 10 image frames, and the road scene type of 6 of the image frames is a vehicle being on a road, then the vehicle being on a road is determined to be the road scene type corresponding to the video data.
[0065] In the disclosed embodiment, road scene types can be divided into three types, road, intersection and roundabout; when the road scene type is that the vehicle is on the road, execute step S340 or S350; when the road scene type is that the vehicle is at an intersection or a roundabout, execute steps S360-S380 or S360-S370-S390.
[0066] S340: If the road scene type is that the vehicle is on the road, and the number of target image frames is greater than or equal to a second preset threshold, it is determined that the video data cannot be used for map drawing.
[0067] S350: If the road scene type is that the vehicle is on the road, and the number of target image frames is less than a second preset threshold, determine that the video data is used for map drawing.
[0068] In an embodiment of the present disclosure, when the electronic device determines that the road scene type corresponding to the video data is a vehicle on a road, it determines whether the number of target image frames is greater than or equal to a second preset threshold value. When the number of target image frames is greater than or equal to the second preset threshold value, it determines that the video data is incomplete, and further determines that the video data cannot be used for map drawing. When the number of target image frames is less than the second preset threshold value, it determines that the video data can be used for map drawing.
[0069] Exemplarily, the second preset threshold is the number of target image frames in the preset video data or the ratio of the preset target image frames to the total image frames in the video data. For example, if the video data contains 50 frames, the second preset threshold may be 5 frames. When the number of target image frames in the video data is greater than or equal to 5 frames, it is determined that the video data cannot be used for map drawing. When the number of target image frames in the video data is less than 5 frames, it is determined that the video data can be used for map drawing.
[0070] In the disclosed embodiment, when the road scene type corresponding to the video data is that the vehicle is on the road, it is possible to determine whether the video data can be used for map drawing based on whether the number of target image frames is greater than or equal to a second preset threshold, thereby improving the recognition efficiency of the video data while improving the accuracy of video data recognition.
[0071] S360: If the road scene type is that the vehicle is at an intersection or a roundabout, and the number of target image frames is at least two, determine a first position and a second position of the vehicle in any two target image frames, respectively.
[0072] S370: Calculate the distance between the first position and the second position.
[0073] S380: When the distance is greater than or equal to a third preset threshold, it is determined that the video data cannot be used for map drawing.
[0074] S390: When the distance is less than a third preset threshold, determine that the video data is used for map drawing.
[0075] In the disclosed embodiment, when the electronic device determines that the road scene type corresponding to the video data is that the vehicle is at an intersection or a roundabout and the number of target image frames is at least two, it is necessary to determine the first position and the second position of the vehicle in any two target image frames, respectively, and calculate the distance between the first position and the second position. When the distance is greater than or equal to the third preset threshold, it is determined that the video data is incomplete, and then the video data cannot be used for map drawing. When the distance is less than the third preset threshold, it is determined that the video data is used for map drawing. It should be noted that if the number of target image frames is greater than two, it is necessary to traverse various combinations consisting of two different target image frames, and then determine the first position and the second position of the vehicle in each combination in the two target image frames. If there is any combination in which the distance between the first position and the second position is greater than or equal to the third preset threshold, it is determined that the video data cannot be used for map drawing. If the distance between the first position and the second position in each combination is less than the third preset threshold, it is determined that the video data is used for map drawing.
[0076] Exemplarily, assuming that the number of target image frames is two, the third preset threshold is 10m, and the two target image frames are the first target image frame and the second target image frame, when the distance between the position of the vehicle in the first target image frame (such as the first position) and the position of the vehicle in the second target image frame (such as the second position) is greater than or equal to 10m, it is determined that the video data cannot be used for map drawing, and when the distance between the position of the vehicle in the first target image frame and the position of the vehicle in the second target image frame is less than 10m, it is determined that the video data can be used for map drawing.
[0077] In the disclosed embodiment, when the road scene type corresponding to the video data is that the vehicle is at an intersection or a roundabout, it is possible to determine whether the video data can be used for map drawing based on whether the distance between the first position and the second position of the vehicle in any two target image frames is greater than or equal to a third preset threshold, thereby improving the recognition efficiency of the video data and the accuracy of the video data recognition.
[0078] In some embodiments of the present disclosure, the acquired video data with data annotation includes multiple video data, and the multiple video data are video data of the same road scene type collected in different time periods. The video data are identified according to the road scene type and the number of target image frames corresponding to the video data to determine whether the video data can be used for map drawing. The method may specifically include: if the road scene type is that the vehicle is at an intersection or a roundabout, for each video data in the multiple video data, determining whether the video data can be used for map drawing according to the number of target image frames; determining the target number of video data that can be used for map drawing in the multiple video data; calculating the ratio of the target number to the total number of the multiple video data; when the ratio is greater than or equal to a fourth preset threshold, determining that the multiple video data cannot be used for map drawing; and when the ratio is less than the fourth preset threshold, determining that the multiple video data can be used for map drawing.
[0079] Exemplarily, the multiple video data are respectively obtained by annotating data on videos of vehicles at intersections or roundabouts collected at different time periods, wherein the videos collected may be videos of vehicles at the same intersection or roundabout at different time periods.
[0080] For example, the fourth preset threshold is 30%. Among multiple video data, for each video data, whether each video data can be used for map drawing is determined according to the number of target video frames in each video data, and then the number of video data that cannot be used for map drawing is determined. When the ratio of the number of video data that cannot be used for map drawing to the total number of multiple video data is greater than or equal to the fourth preset threshold, it is determined that all of the multiple video data cannot be used for map drawing. When the ratio of the number of video data that cannot be used for map drawing to the total number of multiple video data is less than the fourth preset threshold, it is determined that all of the multiple video data can be used for map drawing.
[0081] In the embodiment of the present disclosure, when there are multiple video data, it is possible to determine whether the multiple video data can be used for map drawing, thereby improving the accuracy of video data recognition.
[0082] In the disclosed embodiment, different video data recognition rules can be set according to different road scene types corresponding to the video data to adapt to different road scene types, thereby further improving the accuracy of video data recognition.
[0083] On the basis of the above-mentioned embodiments of the present disclosure, the video data identification method may further include: when the video data cannot be used for map drawing, in response to a recompile operation, re-labeling the video data that has been labeled.
[0084] In the embodiment of the present disclosure, recompiling can be understood as a process of re-labeling data of the collected video.
[0085] In an embodiment of the present disclosure, when video data cannot be used for map drawing, it is determined whether to perform a recompilation operation. When it is determined to perform a recompilation operation, the video data that has been marked is marked again.
[0086] Figure 4 is a flowchart of a method for determining a recompile operation provided by an embodiment of the present disclosure. In some embodiments of the present disclosure, for example Figure 4 As shown, determining whether to perform a recompilation operation may specifically include: when the video data that has been data-annotated cannot be used for map drawing, performing data differential processing on the video data and the preset electronic map data to determine whether there is a preset road scene type in the video data; when it is determined that there is no preset road scene type, marking the video data that has been data-annotated and pushing it; when it is determined that there is a preset road scene type but no annotation data, performing data differential processing on the video data and the collected original video to determine whether the video data contains a saturated video corresponding to the road scene type; when it is determined that the saturated video corresponding to the road scene type is not included, re-collecting the video corresponding to the road scene type, determining to perform a recompilation operation, and then performing a recompilation operation based on the re-collected video; when it is determined that the saturated video corresponding to the road scene type is included, identifying the annotation data in the video data to determine whether there is data missing annotation for the video in the video data; if the judgment result is that there is data missing annotation, determining to perform a recompilation operation; if the judgment result is that there is data missing annotation, marking the video data that has been data-annotated and pushing it.
[0087] When it is determined that the video data does not contain saturated video corresponding to the road scene type, the scope of re-collection is determined according to the unique identifier corresponding to the video data, and a re-collection task is established based on the re-collection scope, and then video collection is performed according to the re-collection task. When there are multiple re-collection tasks, the batch identifier corresponding to the re-collection task is determined according to the unique identifier corresponding to the re-collection task, and multiple re-collection tasks with the same batch identifier are used as target collection tasks to perform video collection, thereby facilitating the management and execution of re-collection tasks.
[0088] In the embodiment of the present disclosure, when recompiling the video data that has been labeled, when the video data that has been labeled includes multiple road scene types, recompiling is performed according to the video data corresponding to each road scene type.
[0089] Exemplarily, two types of road scenes are included in one video data, such as a vehicle on the road and a vehicle at an intersection. In this case, the video data is divided into first video data of the vehicle on the road and second video data of the vehicle at the intersection, and the first video data and the second video data are recompiled respectively.
[0090] In some other embodiments of the present disclosure, determining whether to perform a recompile operation may specifically include: determining to perform a recompile operation in response to a compilation algorithm version update operation and / or a annotation specification change operation.
[0091] Changes in marking specifications can be understood as changes in marking methods, marking labels, etc. For example, the original marking specifications only had one label for intersections. After the marking specifications are changed, the labels for intersections are subdivided into large intersections and small intersections to improve the accuracy of marking.
[0092] Specifically, when the compilation algorithm version is updated, the video data that has been labeled needs to be recompiled. When the labeling specification is changed, in order to ensure the real-time and accuracy of the labeled data, the video data that has been labeled needs to be recompiled.
[0093] In the disclosed embodiment, when the video data that has been data-labeled is re-labeled in response to the recompilation operation, the automated quality inspection operation is started, and when the result of the automated quality inspection is passed, the re-labeled video data is directly pushed; when the result of the automated quality inspection is failed, the manual quality inspection operation is started, and the manual quality inspection is performed by random sampling, and when the result of the manual quality inspection is passed, the re-labeled video data is directly pushed; when the result of the manual quality inspection is failed, in response to the failure of the manual quality inspection, the video data that failed the manual quality inspection is deleted, and only the video data that passed the manual quality inspection is retained, and then the qualified video data in the re-labeled video data is pushed.
[0094] In the disclosed embodiment, when the video data that has been annotated cannot be used for map drawing, it can be determined whether to perform a recompilation operation. When it is determined to perform a recompilation operation, the video data that has been annotated is annotated again according to a preset annotation algorithm, so as to reduce the data loss of the collected video, improve the utilization rate of the video, and realize the rapid update of the annotated video data.
[0095] Figure 5: is a schematic diagram of the structure of a video data recognition device provided in an embodiment of the present disclosure. The video data recognition device in the embodiment of the present disclosure can be set in an electronic device, and the electronic device can be a server or a terminal or a server cluster, wherein the terminal can specifically include a mobile phone, a computer or a tablet computer, a vehicle-mounted terminal, or any device that can be used to process the video data recognition method, etc., which is not limited here.
[0096] like Figure 5 As shown, the video data identification device 500 may include a data acquisition module 510 , a first determination module 520 and a second determination module 530 .
[0097] The data acquisition module 510 may be used to acquire video data that has been annotated.
[0098] The first determination module 520 may be configured to determine a target image frame in the video data according to the annotation data in the video data, wherein the target image frame is an image frame that lacks annotation data corresponding to a preset number of viewing angles.
[0099] The second determination module 530 may be used to identify the video data according to the number of target image frames, and determine whether the video data is target video data, where the target video data is video data that cannot be used for map drawing.
[0100] In an embodiment of the present disclosure, it is possible to obtain video data that has been annotated with data, and determine at least one target image frame in the video data based on the annotation data of each image frame in the video data that has been annotated with data, wherein the target image frame is an image frame whose number of viewing angles included in the annotation data is less than or equal to a first preset threshold value. Then, based on the number of target image frames, the video data is identified to determine whether the video data can be used for map drawing. Thus, it is possible to determine the target image frame by determining whether the number of viewing angles included in the annotation data of each image frame in the video data is less than or equal to a first preset threshold value, and then determine whether the video data can be used for map drawing based on the number of target image frames in the video data. There is no need for manual judgment on whether the video data can be used for map drawing, thereby reducing labor costs and improving the recognition efficiency of video data.
[0101] In some embodiments of the present disclosure, the first determination module 520 may include a viewing angle determination unit 5201 and an image frame determination unit 5202 .
[0102] The viewing angle determination unit 5201 may be configured to determine, for each image frame, the number of viewing angles included in the annotation data corresponding to each image frame.
[0103] The image frame determining unit 5202 may be configured to determine an image frame as a target image frame when the number of viewing angles is less than or equal to a first preset threshold.
[0104] In some embodiments of the present disclosure, the second determination module 530 may be specifically used to identify the video data according to the road scene type and the number of target image frames corresponding to the video data, and determine whether the video data can be used for map drawing.
[0105] In some embodiments of the present disclosure, the second determination module 530 may include a first determination unit 5301 and a second determination unit 5302 .
[0106] The first determination unit 5301 may be configured to determine that the video data cannot be used for map drawing if the road scene type is that the vehicle is on the road and the number of target image frames is greater than or equal to a second preset threshold.
[0107] The second determination unit 5302 may be configured to determine that the video data is used for map drawing if the road scene type is that the vehicle is on the road and the number of target image frames is less than a second preset threshold.
[0108] In some embodiments of the present disclosure, the second determination module 530 may include a third determination unit 5303 , a distance calculation unit 5304 , a fourth determination unit 5305 and a fifth determination unit 5306 .
[0109] The third determination unit 5303 can be used to determine the first position and the second position of the vehicle in any two target image frames respectively if the road scene type is that the vehicle is at an intersection or a roundabout and the number of target image frames is at least two.
[0110] The distance calculation unit 5304 may be configured to calculate the distance between the first position and the second position.
[0111] The fourth determination unit 5305 may be configured to determine that the video data cannot be used for map drawing when the distance is greater than or equal to a third preset threshold.
[0112] The fifth determination unit 5306 may be configured to determine that the video data is used for map drawing when the distance is less than a third preset threshold.
[0113] In some embodiments of the present disclosure, multiple video data are included, and the multiple video data are video data of the same road scene type collected in different time periods.
[0114] The second determination module 530 may include a sixth determination unit 5307 , a seventh determination unit 5308 , a ratio calculation unit 5309 , an eighth determination unit 5310 , and a ninth determination unit 5311 .
[0115] The sixth determination unit 5307 may be used to determine, for each of the multiple video data, whether the video data can be used for map drawing based on the number of target image frames if the road scene type is that the vehicle is at an intersection or a roundabout.
[0116] The seventh determination unit 5308 may be configured to determine a target number of video data that can be used for map drawing among the plurality of video data.
[0117] The ratio calculation unit 5309 may be configured to calculate a ratio of a target quantity to a total quantity of multiple video data.
[0118] The eighth determination unit 5310 may be configured to determine that the plurality of video data cannot be used for map drawing when the ratio is greater than or equal to a fourth preset threshold.
[0119] The ninth determination unit 5311 may be configured to determine that the plurality of video data can be used for map drawing when the ratio is less than a fourth preset threshold.
[0120] In some embodiments of the present disclosure, the video data identification device 500 may further include a data annotation module 540 .
[0121] The data annotation module 540 may be used to re-annotate the video data that has been annotated with data in response to a recompile operation when the video data cannot be used for map drawing.
[0122] It should be noted that Figure 5 The video data identification device 500 shown can execute each step in the above method embodiment and realize each process and effect in the above method embodiment, which will not be described in detail here.
[0123] Figure 6 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown.
[0124] In the disclosed embodiment, Figure 6 The electronic device shown may be a server or a terminal or a server cluster, wherein the terminal may specifically include a mobile phone, a computer or a tablet computer, a vehicle-mounted terminal, or any device that can be used for the video data recognition method, etc., and is not limited here.
[0125] like Figure 6 As shown, the electronic device may include a processor 610 and a memory 620 storing computer program instructions.
[0126] Specifically, the processor 610 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.
[0127] The memory 620 may include a large capacity memory for information or instructions. By way of example and not limitation, the memory 620 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a tape, or a universal serial bus (USB) drive or a combination of two or more of these. Where appropriate, the memory 620 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 620 may be inside or outside the integrated gateway device. In a particular embodiment, the memory 620 is a non-volatile solid-state memory. In a particular embodiment, the memory 620 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (Electrically Erasable Programmable ROM, EEPROM), an electrically rewritable ROM (EAROM) or a flash memory, or a combination of two or more of these.
[0128] The processor 610 reads and executes the computer program instructions stored in the memory 620 to perform the steps of the video data identification method provided in the embodiment of the present disclosure.
[0129] In one example, the electronic device may further include a transceiver 630 and a bus 640. Figure 6 As shown, the processor 610, the memory 620 and the transceiver 630 are connected via a bus 640 and communicate with each other.
[0130] The bus 640 includes hardware, software, or both. For example, but not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel 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 Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, the bus 640 may include one or more buses.
[0131] The embodiments of the present disclosure further provide a computer-readable storage medium, which may store a computer program. When the computer program is executed by a processor, the processor implements the video data recognition method provided by the embodiments of the present disclosure.
[0132] The above-mentioned storage medium may, for example, include a memory 620 of computer program instructions, and the above-mentioned instructions may be executed by the processor 610 of the electronic device to complete the video data recognition method provided by the embodiment of the present disclosure. Optionally, the storage medium may be a non-temporary computer-readable storage medium, for example, the non-temporary computer-readable storage medium may be a ROM, a random access memory (Random Access Memory, RAM), a compact disc read-only memory (Compact Disc ROM, CD-ROM), a magnetic tape, a floppy disk, and an optical data storage device, etc.
[0133] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0134] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A video data recognition method, It is characterized in that The method comprises: Obtain video data that has been annotated; Determine, according to the annotated data of each image frame in the video data, at least one target image frame in the video data, the target image frame being an image frame whose number of viewing angles included in the annotated data is less than or equal to a first preset threshold; The video data is identified according to the number of the target image frames to determine whether the video data can be used for map drawing.
2. The method according to claim 1, It is characterized in that The step of determining at least one target image frame in the video data according to the labeled data of each image frame in the video data comprises: For each of the image frames, determining the number of viewing angles included in the annotation data corresponding to each of the image frames; When the number of viewing angles is less than or equal to the first preset threshold, the image frame is determined to be a target image frame.
3. The method according to claim 1, It is characterized in that The step of identifying the video data according to the number of the target image frames and determining that the video data can be used for map drawing includes: The video data is identified according to the road scene type corresponding to the video data and the number of the target image frames, and it is determined that the video data can be used for map drawing.
4. The method according to claim 3, It is characterized in that The step of identifying the video data according to the road scene type corresponding to the video data and the number of the target image frames, and determining that the video data can be used for map drawing, includes: If the road scene type is that the vehicle is on the road, and the number of the target image frames is greater than or equal to a second preset threshold, determining that the video data cannot be used for map drawing; If the road scene type is that the vehicle is on the road and the number of the target image frames is less than the second preset threshold, it is determined that the video data is used for map drawing.
5. The method according to claim 3, It is characterized in that The step of identifying the video data according to the road scene type corresponding to the video data and the number of the target image frames, and determining that the video data can be used for map drawing, includes: If the road scene type is that the vehicle is at an intersection or a roundabout, and the number of the target image frames is at least two, determining a first position and a second position of the vehicle in any two target image frames respectively; calculating a distance between the first position and the second position; When the distance is greater than or equal to a third preset threshold, determining that the video data cannot be used for map drawing; When the distance is less than the third preset threshold, it is determined that the video data is used for map drawing.
6. The method according to claim 3, It is characterized in that The plurality of video data are video data of the same road scene type collected at different time periods; The step of identifying the video data according to the road scene type corresponding to the video data and the number of the target image frames, and determining whether the video data can be used for map drawing, includes: If the road scene type is that the vehicle is at an intersection or a roundabout, for each of the plurality of video data, determining whether the video data can be used for map drawing according to the number of the target image frames; determining a target number of video data from the plurality of video data that can be used for mapping; Calculating a ratio of the target quantity to a total quantity of the plurality of video data; When the ratio is greater than or equal to a fourth preset threshold, it is determined that the plurality of video data cannot be used for map drawing; When the ratio is smaller than the fourth preset threshold, it is determined that the plurality of video data can be used for map drawing.
7. The method according to claim 1, It is characterized in that The method further comprises: When the video data cannot be used for map drawing, the video data that has been marked with data is marked again in response to a recompiling operation.
8. A video data recognition device, It is characterized in that include: A data acquisition module, used to acquire video data that has been annotated; A first determination module, configured to determine a target image frame in the video data according to the annotation data in the video data, wherein the target image frame is an image frame that lacks annotation data corresponding to a preset number of viewing angles; The second determination module is used to identify the video data according to the number of the target image frames, and determine whether the video data is target video data, and the target video data is video data that cannot be used for map drawing.
9. An electronic device, It is characterized in that include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and is configured to be executed by the processor to implement the method according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Panoramic vision SLAM method based on multi-camera cooperation
CN110070615A
Special transport vehicle monitoring method and device, medium and electronic equipment
CN111523527A
Map data updating method and device, storage medium and electronic equipment
CN112269851A
Method and device for identifying landmark in video clip, equipment and storage medium
CN113963295A
Multi-view video transmission method and device, processing equipment, display equipment and medium
CN114222092A
